NOTE: Some content may not display correctly, including tables and figures. See PDF for full details.

Preparation of National 

Electrification Rollout 

Plan and Financing Prospectus Final Report 

11 April 2017 

Submitted to DPE and the World Bank by: 

The Earth Institute 

Economic Consulting Associates 

Sustainable Engineering Lab                                                                                  Economic Consulting Associates Limited 

School of Engineering and Applied Sciences and                                                     41 Lonsdale Road, London NW6 6RA, UK 

Earth Institute, Columbia University tel: +44 20 7604 4546 tel: +1-212-854-2956 fax: +44 20 7604 4547 [email protected]  www.eca-uk.com 

Contents  

Contents

Abbreviations and acronyms                                                                                           vii

Executive summary                                                                                                                1

Introduction to this report                                                                                                  11

PART 1: GEOSPATIAL LEAST-COST NATIONAL ELECTRIFICATION 

PLAN                                                                                                                                      12

1                   Introduction           13

2                   Data collection and preparation            14

2.1              Geo-referenced PPL distribution system map 14

2.2              Geo-located demands       18

2.3              Technical and cost parameters    22

2.4              The need for a geospatial planning platform    32

3                   Estimate of current grid access 33

4                   Final Model Results         35

4.1              Final results for the “Base Case”             36

4.2              Costs of grid extension per decile and household (“cost build-up”)            42

4.3              Results of a high cost analysis reflecting recent project costs             46

4.4              Potential approaches to cost reduction 50

PART 2: IMPLEMENTATION OF THE LEAST-COST NATIONAL 

ELECTRIFICATION PLAN                                                                                               53

5                   Institutional framework 54

5.1              Introduction           54

5.2              Existing framework           55

5.3              Evaluation of options        67

5.4              Summary of recommendations   83

6                   Funding mechanisms       106

6.1              Introduction           106

6.2              Existing mechanisms        107

6.3              Evaluation of options        115

6.4  Summary of recommendations                                                                                 120

7  Environmental and social considerations                                                                124 

7.1  Introduction                                                                                                                  124 

7.2  Description of the different options for delivery access to electricity                 125 

7.3  Potential environmental and social impacts                                                           128 

7.4  Legal and regulatory context                                                                                     132 

7.5  Assessment of institutional capacity                                                                        155 

8  Financing plan                                                                                                               158 

8.1  Introduction                                                                                                                  158 

8.2  Electrification activities under NEROP                                                                    159 

8.3  The investment requirements of NEROP                                                                164 

8.4  Availability of financing for NEROP                                                                        170 

8.5  NEROP financing plan                                                                                               175 

ANNEXES                                                                                                                           183 

A1  All Geospatial Modelling Parameters                                                                    184 

A2  Assessment of Inter-Household Distance                                                             187 

A3  30% Correction Factor for MV Length                                                                   191 

A4  Geospatial Plan: Sensitivity Analysis                                                                    192 

A5  International case studies on electrification                                                         195 

A5.1  Centralised approaches                                                                                           195 

A5.2  Decentralised approaches                                                                                       198 

A5.3  Hybrid approaches                                                                                                   200 

A5.4  Community approaches                                                                                          202 

A5.5  Funding                                                                                                                      203 

Tables, figures and boxes

Tables, figures and boxes 

Tables  

Figure 39 Annual capex and electrification rate – higher unit costs                                     179

Figure 40: Open wire plus ABC approximation of inter-household spacing                        188 

Figure 41: Area-based approximation of inter-household spacing                                       189 

Figure 42: Single line approximation of inter-household spacing                                         189 

Figure 43: Results of visual review of satellite imagery for household spacing.                  190 

Figure 44: 30% "correction factor" addresses the complex pathways between locations      191 

Boxes 

Box 1 Electrification planning under NEROP                                                                         96

Box 2 Calculating an affordable tariff                                                                                    175

Box 3 Utility led electrification in Lao PDR                                                                           195

Box 4 State led electrification in Vietnam                                                                              196

Box 5 Rural electrification office in Morocco                                                                         197

Box 6 Private sector participation in Chile                                                                            198

Box 7 Decentralised electrification in Burkina Faso                                                              199 

Box 8 Qualified third party model in the Philippines                                                           200 

Box 9 Rural electrification in Ethiopia                                                                                   201

Box 10 Community involvement in Bangladesh                                                                   202

Box 11 Wholesale power subsidies in Thailand                                                                    203

Box 12 Franchise areas in the Philippines                                                                              203

Box 13 Capital subsidies in Peru                                                                                            204

Box 14 Operational subsidies in Peru                                                                                    205

PNG-NEROP-FinalReport-2017-04-11.docx 12/04/2017  

Abbreviations and acronyms  

Abbreviations and acronyms 

Abbreviations and acronyms  

Executive summary 

About this report 

This is the final report for geospatial least-cost modelling by the Columbia University consulting team for the National Electrification Roll-Out Plan (NEROP) for Papua New Guinea.  

Part 1 of this report describes the results of geospatial cost and technical modelling as conducted by Sustainable Engineering Lab at the Earth Institute at Columbia University (SEL/EI).  

Part 2 of this report details the implementation of NEROP, including the institutional framework, funding mechanisms, environmental and social safeguards, and financial plan, as prepared by Economic Consulting Associates (ECA). 

Part 1: Geospatial Least-Cost National Electrification Plan 

Preparation of the geospatial plan 

The final modelling and analysis work took place in June and July, 2016, following completion of the medium voltage grid infrastructure mapping effort conducted by PNG Power (with support from the Columbia team) in March, 2016, and the presentation of preliminary results at the June 15, 2016, NEROP workshop in Port Moresby, Papua New Guinea. 

This Final Report describes final results of algorithmic modelling performed by the team from the SEL/EI.The approach used in this modelling has been to:  a) employ the bestavailable data for electricity demands (primarily settlements, education and health facilities) and the electricity grid network, b) compare the costs of electrification of unserved populations by grid, mini-grid and off-grid/solar technologies (informed by local and international information on costs and technical standards), and c) report on the recommended technology option, as well as the cost and technical capacity and other details, for each location.  

Key insights from the geospatial modelling 

The following summary provides the main insights gained from this geospatial least-cost modelling work:  

        PPL’s smartphone GPS mapping of the utility’s distribution grid was successful. The utility’s map of MV grid lines and related equipment (transformers, substations and switches) provides a sound basis for geospatial planning.  Used in combination with other geospatial data (such as NSO census data and geolocated social infrastructure points) it has allowed an estimate of the national population within range of low-cost grid access, as well as least-cost modelling for future MV extensions and non-grid electrification options.  

Part 2: Implementation of the Least-Cost National Electrification Plan 

Institutional framework 

The current institutional framework 

The current institutional framework for implementing rural electrification in PNG is largely defined in the Electricity Industry Policy (EIP). Although the EIP provides for private sector involvement in electricity supply, PPL has led almost all electrification projects to-date. Because electrification is a loss-making activity for PPL and PPL faces constraints on its financial and technical capacity, relatively few electrification projects have been implemented historically (approximately 2,100 connections per year). 

The key participants in rural electrification at present are summarised in the figure below. 

Key electrification challenges and requirements 

International experience shows that the exact institutional structure is not critical to a country’s success in implementing an electrification program. What is more important is that the chosen structure suits the specific circumstances of the particular country. Nevertheless, there are a few fundamental challenges that all electrification programs face, including: 

There are also a few critical lessons that can be drawn from international experience: 

        Clear performance indicators and incentives: Entities must have clear targets for delivery and incentives to meet these. 

        Ring-fence from commercial activities: Given that electrification is generally loss-making, particularly mini-grids. 

        Avoid creating excessive costs: Having multiple small entities reduces economies of scale and increases administrative and regulatory costs. 

        Ensure sustainability: Entities responsible for operating and maintaining systems must have sufficient technical and financial capacity to do so. 

Recommended institutional framework for implementing NEROP 

We recommend that PNG adopt a hybrid approach to implementing NEROP, with PPL responsible for grid extensions and the private sector responsible for establishing new offgrid solutions, including mini-grids. If there is insufficient interest from the private sector in establishing mini-grids, the obligation falls back on PPL.  

This approach combines the key advantages of centralised implementation – economies of scale, use of existing technical and commercial expertise, the ability to directly control the speed of the rollout – with the key advantages of a decentralised model – utilisation of private sector capital and expertise, overcoming capacity constraints on the utility, reducing costs through competition, independence from political interference.  

More specifically we recommend that: 

We recommend that the lead primary policy making agency (DPE) define electrification targets for both PPL and OGEA, including selecting and prioritising the areas to be served (by either PPL-led grid extension or OGEA-led mini-grids), administer the electrification funds and ensure that available funding aligns with the electrification targets, and encourage and mobilise the private sector, including involving communities where possible. 

DPE will need to work closely with a number of other government departments, in particular with the Department of Provincial and Local Government Affairs (DPLGA) to identify/prioritise communities for electrification and to allocate SIP funding to electrification projects and the Department of National Planning and Monitoring (DNPM) to monitor progress implementing NEROP. 

Our recommended institutional framework is summarised in the figure below. 

Technical assistance required 

While many activities required to implement NEROP are not entirely new, the large scale and aggressive pace of NEROP suggest that technical capacity, staffing and financial resources will all need to be increased, both to ensure the best possible execution of NEROP, and to avoid overwhelming existing resources. In addition, NEROP will involve the creation of a new agency, OGEA, to manage off-grid electrification and whose capacity will largely need to be built from scratch. Costs for this technical assistance are estimated below. 

Source: Consultants 

While the estimated budget may appear high, this is driven by our assessment of what is necessary to ensure that NEROP is implemented on schedule and in a manner that fully complies with the various procurement and safeguards policies of the Government and development partners from its inception. 

Funding mechanisms 

The current funding mechanisms 

PPL currently relies on cross-subsidies, by applying a uniform tariff, to operate expensive diesel mini-grid systems. PPL has not however increased tariffs in recent years and its financial capacity is limited, which makes funding further electrification projects difficult.  

The EIP envisages that future grid extensions to unprofitable areas will be funded through up-front subsidies and grid-specific tariffs. One key mechanism for the Government to provide up-front subsidies is the national CSO framework, however it has yet been fully implemented and therefore no centralised funding is available at present. Similarly, the Electricity Trust Fund envisaged in the EIP has not been established. 

International experience with funding electrification 

No successful rural electrification program worldwide has functioned without some form of subsidy, although the exact form of subsidy is not critical. What matters most is costrecovery –- operators must be able to recover the full costs of supply through a combination of tariffs and subsidies.  

Capital cost subsidies are the primary mechanism for funding rural electrification programs worldwide. By reducing the installation cost of the generators and distribution grid, either through grant funding or concessionary loans, many schemes are then able to charge affordable tariffs that fully cover the remaining costs of operating the system. 

The cost of operating mini-grids in PNG (including capital replacement) is likely to exceed $0.50/kWh in many cases, which may limit consumption for many rural households. While this can likely be reduced in some cases by reducing service standards and implementing off-grid solutions such as solar home systems, we still expect that capital, operational, and connection subsidies will be needed in many cases to ensure that PNG operators can fully recover their costs. 

Recommended funding mechanisms 

Our key recommendations for funding NEROP include: 

an operational subsidy if the cost-recovery tariff is higher than the affordability cap set by the regulator. This arrangement effectively puts in place a crosssubsidy from main grid customers to mini-grid customers. 

 The charges for connecting to the grid or mini-grids should be kept below cost, particularly for rural households whom have limited ability to make a large one-off payment. 

Our recommended funding mechanisms are summarised in the figure below. 

Financing plan 

Current electrification 

Approximately 12.5% of households in PNG are currently connected to grid-based power supply. Geospatial analysis of population centres and the existing grid shows that approximately 6.5% of households are within 1km of an existing grid transformer but not connected. The remaining 81% of households are further than 1 km from the grid. 

Investment requirement 

The total cost of achieving 70% electrification by 2030 is likely to be around US$1.8 billion. This amount hinges on the extent to which the costs of distribution line can be brought down through economies of scale and improved procurement practices. At PPL’s current costs, the cost would be around US$3.6 billion. These estimates include both grid costs and the cost of off-grid solutions, although the costs of off-grids are more uncertain as it depends on the delivery model chosen (solar home system, diesel mini-grid, hybrid system etc.).  

This investment cost equates to an average cost per year of approximately US$126 million (or US$260 million at PPL’s current costs) from 2017 to 2030, and US$1,318 (or US$2,719) per household connected.  

The annual cost (assuming unit costs that are approximately 40% lower than PPL’s current costs) is summarised in the figure below. 

Financing investments 

To finance NEROP, the Government will be required to contribute approximately US$22m per year on average (less in the initial years, more in the later years). This assumes that connection charges would finance US$14m per year on average and development partners the remaining US$90m per year in the form of concessionary loans. This assumes that development partners will be willing to contribute 70% of the total cost (80% of the financing gap), which is likely the upper bound on what can realistically be achieved.  

The costs of financing NEROP in the first five years are summarised in the table below. It assumes donor funding is not available until 2018.  

Grid levy to cover recurring costs 

Current tariffs will not be sufficient to cover the costs of operating new schemes. Many offgrid schemes may not even cover their costs of supply even at maximum affordable tariff levels (US$0.38/kWh). To fund this shortfall in operating/recurring costs, a levy will need to be collected from all grid-connected customers, which may reach US$0.04/kWh by 2030. 

Introduction to this report  

Introduction to this report 

This is the Final Report for the National Electrification Roll-Out Plan (NEROP) for Papua New Guinea. The document consists of two parts: 

This report presents the final results for the NEROP project. The Investment Prospectus is provided separately.  

Part 1: Geospatial Least-Cost National Electrification Plan  

PART 1: GEOSPATIAL LEAST-COST NATIONAL ELECTRIFICATION PLAN

Introduction  

1 Introduction 

This document presents the final results for the geo-spatial technical and cost analysis for the project “Preparation of National Electrification Least Cost Geospatial Rollout Plan (grid and off-grid) and Sector-wide Investment Financing Prospectus,” implemented by the Papua New Guinea Department of Petroleum and Energy (DPE). These results have been prepared by Columbia University – specifically the Sustainable Engineering Lab, directed by Prof. Vijay Modi, which is part of the Earth Institute at Columbia University.  This technical and cost modelling is the foundation for the geospatial least cost national electrification rollout plan for the Government of Papua New Guinea, with a target of achieving 70% electricity access nationwide by 2030. 

This work emerges from the country’s national electrification goals and planning process, as outlined in the Government’s Electricity Industry Policy (EIP).  Recognizing electricity as a key enabler for achieving economic growth, widely shared prosperity, and modernization, the Government of Papua New Guinea (GoPNG) has prioritized increasing electricity access as an important element of the country’s Vision 2050. To address this challenge, the EIP aims to put in place the institutions, processes, and mechanisms to enable achievement of the national goals of reliable supply and affordable access for all citizens in an efficient and equitable manner.  The EIP established the Electricity Management Committee (EMC) as the overarching coordinating body to implement and achieve the objectives set out in the EIP, with the Secretary of the Department of Petroleum and Energy (DPE) as the EMC’s chair.  A major policy initiative launched under the EIP is the preparation of the National Electrification Rollout Plan (NEROP) to scale up access nationwide.  The geo-spatial leastcost national electrification plan that immediately follows is one of the key components of NEROP.  This geo-spatial analysis also forms the basis for other key NEROP components, including the parallel effort by Economic Consulting Associates focused on preparation of the Prospectus for Investment Financing for the first 5 years of the project, also described in this report.  

The scope of work broadly encompasses: geo-located electricity demand projections; evaluation and comparison of grid and non-grid electrification technologies for each demand location; and creation of a least-cost GIS electrification roll-out plan with attention to the plan’s sensitivity to key variables and policies.  These have been performed by an open-source, web-based software system for electrification cost modelling designed and used by SEL/EI in New York City, called “NetworkPlanner.” Results of this work were used by another member of this consulting consortium (ECA) as a key input for that firm’s preparation of preliminary results for an investment plan, as well as to inform investigations of institutional, regulatory, environmental and social aspects of this project. All results of 

this work were presented at the NEROP workshop in Port Moresby, Papua New Guinea, on 

August 17, 2016.                                                 

2 Data collection and preparation 

Initial efforts to collect and prepare data were described in Inception Report, including: collection of geo-located populated places and other electricity demand locations, initial training for mapping of medium voltage grid infrastructure, and collection of information regarding technical and cost parameters for modelling. 

A very brief review of the main components of the SEL electrification planning approach helps to put this data collection work in context. The SEL model relies upon three main types of information: (i) geolocated demand points (populated places and social infrastructure), (ii) a map of medium-voltage grid infrastructure (MV lines and transformers), and (iii) 75-80 modelling parameters related to settlement patterns, technical aspects and costs of grid and non-grid electrification projects and technologies. The model employs this information in three basic steps. First, the model computes future electricity demand for all locations in the system along with the costs to meet this demand using three electrification technologies (grid, diesel mini-grid and solar systems); second, the model uses a combination of location and cost data to propose a grid network and non-grid system locations; and, third, the model performs a simple computation comparing initial investment costs vs. electricity delivered to each grid location to prioritize the grid extension to each location (prioritization of non-grid systems is not specified in this step).  

Building upon the description provided in the Inception Report, the following section details subsequent efforts to complete, refine, vet or otherwise improve the underlying geospatial dataset. 

2.1                Geo-referenced PPL distribution system map 

SEL/EI conducted a week long training for a group of PPL teams during the inception visit in February, 2016, focused on use of GPS-enabled smartphones to map the existing medium voltage (MV) grid infrastructure, including lines, transformers and other equipment. Within a week following the training, PPL had put in place field logistics and vehicles for 5 separate mapping teams. Within about 40 days, PPL successfully mapped a total of approximately 4,100 km of MV lines, comprising 1,880 km of 11 kV lines (primarily in urban areas) and 2,216 km of 22 kV lines (primarily in rural areas). The PPL teams also simultaneously mapped other grid infrastructure and assets like, power plants, injection substations, transformers, switches, isolators and end poles. A summary of the features mapped is presented in Table 1 below (and all of this data is accessible via an online geo-database portal at ppl.gridmaps.org).

Table 1: Summary MV grid infrastructure features mapped by PPL, February, 2016 

The locations and lengths of these lines, identified according to the “Operator” that the field mapping team assigned during the data gathering process, are presented in Figure 1 below. Other features, such as transformers, substations and switches, which were mapped as point locations, are shown in Figure 2 and Figure 3. 

Figure 3: Other grid infrastructure (switches, end poles) 

Some concluding notes on the effort and outcome of the grid mapping program: 

2.2 Geo-located demands 

The most important dataset for geolocated demands was the 2011 Census data obtained from NSO which provided residential demand points in the form of geolocated census units with population. It was noted in the Inception Report that this dataset would need to be cleaned to prepare a standardized national dataset. The key issue with this dataset was the absence of georeferencing coordinates for some of the census units.  

2.2.1 Census Unit records (CUs) without clear geolocation information  

Of the list of about 27,000 Census Unit (CU) records in the 2011 Census data received from NSO, about 22,000 CUs, representing about 6.2 million people, are geolocated (with latitude and longitude coordinates). However, about 5,000 CU records – representing about 1.04 million people, or about 14% of the total national population – lack geographic 

coordinates^[1].On further scrutiny^[2], about 1,100 of the non-geolocated CU records are urban CUs representing about 187,000 people, or about 3% of the national population. The remaining set of CU records are non-geolocated – representing 850,000 people, or 11% of the national population – are identified as rural locations. 

NSO staff described that the CUs that lack latitude/longitude coordinates are new CUs – they did not exist in the prior census (2000) and were added as new records for the 2011 census. These new CUs fall into two groups: a) CUs that have no obvious relationship to the geocoded CU points from the prior census year (2000) and b) those that are clearly split or sub-divided from pre-existing CUs (present in the 2000 census).Figure 4 below shows the distribution of these non-geolocated points at the LLG level nationwide. Of the country’s 326 LLGs, about 189 LLGs, mostly in the northern and highland regions, contain most of the non-geolocated census units. Of these, most of the LLGs (about 146) have less than 20% of their census units non-geolocated, however there are few LLGs (~7) for which almost all CUs are non-geolocated. The presence of non-geolocated CUs is particularly an issue for two new provinces of the Highlands region, Hela (which split from Southern Highlands) and Jiwaka (which split from Western Highlands).The PNG Parliament approved these new provinces in 2009, and they officially came into being on 17 May 2012.^[3]

Figure 4:  Non-geolocated CUs at the LLG level (2011 Census)

To resolve this data gap, these non-geolocated populations were assigned to a geolocated point called a mean spatial center, or “centroid”. Centroids can be created from points, lines or polygons (see Figure 5 below). 

For most non-geolocated points, records for other CUs in the same ward contained geolocation information (latitude / longitude coordinates). For these CU records, the nongeolocated CUs were grouped by ward and their population was aggregated. All geolocated CUs within the same ward were then identified, and a “centroid” for each ward was calculated (see the left-most example in Figure 5 above).^[4] This newly created, geolocated centroid was then used to represent all the non-geolocated points within that specific ward and assigned the sum of the non-geolocated population. 

This method of assigning population to centroids required adaptation for the few nongeolocated census units which did not have corresponding geolocated census units within the same ward. The population values for these CUs were spatially aggregated to the next higher administrative level, the Low Level Government (LLG).Then a centroid was created for the LLG polygons^[5] (see the two right-most examples in Figure 5 above), and the aggregated population was assigned to the centroid point. 

Figure 7: Centroid (red) represents all non-geolocated population for CUs in an LLG 

Finally, about 3,600 CUs representing about 58,500 people, included geographic coordinates that duplicated at least one other CU record. For these cases, CU records were again grouped based on their NSO-assigned geocodes (for wards and other administrative levels) and the population for all census units with the duplicate geographic coordinates were summed up and assigned to one CU point. 

In these ways, all census units with either missing or duplicated geolocation information were assigned some location – often within the same ward, in rare cases within the same LLG – such that the population was preserved throughout the analysis, and the costs for electrification were included, with some approximation, in the geo-spatial least-cost plan. 

2.2.2 Other steps for preparing the geolocated demand dataset 

Other steps in preparation of the demand point dataset are summarized here in brief: 

infrastructure (such as grid lines). This latter component is typically set at 25-30 years based on international accounting approach for major infrastructure investments. These two are combined under a single time horizon parameter in the model by recalculating the population growth rates to yield the same results in the 30-year amortization time frame that they would in the 2016-2030 timeframe that defines the target year. 

2.3 Technical and cost parameters 

It is necessary to establish around 75 parameter values for a successful model run. Using the best possible parameters is essential for making realistic cost and technical estimates that will be credible to the utility, private sector, and other development partners. For this reason, as much as possible, these parameters are gathered locally, with the majority of the information coming from PPL itself.^[6] Others have been obtained from the private sector (primarily from Certway Power) and international comparison. The resulting parameters are listed in Annex A1. The ten “core” parameters presented in Table 2 below are both critical to the overall outcome of the analysis and are also generally applicable to all electrification technologies (grid, mini-grid or off-grid / solar).For roughly half of these “core” parameters, this analysis has assigned a specific value for each location in the dataset, based on geographic data. This data includes growth rates provided by NSO, wealth and poverty data from the World Bank, and other similar sources that provide additional details for how parameters change over the landscape. The key parameters are described in more detail in the following sub-sections. 

Table 2: “Core” model parameters that apply to all electrification technologies. 

2.3.1 Estimating household demand 

Household demand (expressed in kWh per household per year) is arguably the single most important parameter value for the entire modelling exercise. This parameter is critically important in part because it affects overall system sizes and cost, but also because household demand strongly influences the recommended electrification technology reported by the model. In short, over the long-term (15 years or more):higher annual household demand tends to favor grid connectivity as the least-cost option, since the lower recurring costs of grid outweigh the costs of long-lived but relatively expensive grid lines; very low household demand tends to favor solar home systems as the least-cost option, since this system type has relatively low initial costs, and relatively high recurring costs (due to battery replacement, primarily); and intermediate demands tend to be recommended for mini-grids, since these involve moderate investment in lines, but also have fairly high recurring costs (for diesel fuel or battery replacement). 

Household electricity demand can vary with cooking practices, ownership of certain appliances, and others, all of which can change over time and throughout the landscape of a target area. For this project, the first step in estimating household demand was to establish a reasonable range of consumption values spanning rich and poor homes, in urban and rural areas. This was undertaken through a review of PPL billing data gathered from the Port Moresby headquarters and PPL field offices in Kokopo and Madang.^[7] The data obtained is shown in Figure 8 below. The demand range extends from a low of about 25 kWh per month (300 kWh per year) to a medium-high level of perhaps 200 kWh per month (~2,500 kWh per year). The minimum of 300 kWh per year is confirmed by PPL staff who state that a household must purchase at least 15 PGK of power each month to maintain a PPL connection.^[8] The maximum value for this estimated range is harder to clarify, in part because one PPL connection may serve needs beyond a single house (shops, neighboring households).It is also important that this modelling work consider the likely demand of rural and poor residents who are will be the target of the Sustainable Energy for All access program. 

The next step in estimating household demand was to assign specific household demand values within this range to each location in the dataset of census unit points. This was done by dividing locations into urban and rural areas (using ward codes provided by NSO) and then estimating the percentage of poor in each location based on a 2004 World Bank study (see equations below for urban and rural household demand). Specifically, the poverty rate at the province level was used to calculate the percentage of poor and non-poor households in each census unit. Then these percentages were applied to the poor and rich electricity consumption numbers, using different electricity demand ranges for urban and rural areas.  

𝑼𝒓𝒃𝒂𝒏 𝒂𝒏𝒏𝒖𝒂𝒍 𝒉𝒐𝒖𝒔𝒆𝒉𝒐𝒍𝒅 𝒆𝒍𝒆𝒄𝒕𝒓𝒊𝒄𝒊𝒕𝒚 𝒅𝒆𝒎𝒂𝒏𝒅

𝑁𝑜𝑛 𝑝𝑜𝑜𝑟 ℎ𝑜𝑢𝑠𝑒ℎ𝑜𝑙𝑑 𝑛𝑢𝑚𝑏𝑒𝑟×2500 + 𝑃𝑜𝑜𝑟 ℎ𝑜𝑢𝑠𝑒ℎ𝑜𝑙𝑑 𝑛𝑢𝑚𝑏𝑒𝑟×1000 =  

𝑁𝑜𝑛 𝑝𝑜𝑜𝑟 ℎ𝑜𝑢𝑠𝑒ℎ𝑜𝑙𝑑 𝑛𝑢𝑚𝑏𝑒𝑟 + 𝑃𝑜𝑜𝑟 ℎ𝑜𝑢𝑠𝑒ℎ𝑜𝑙𝑑 𝑛𝑢𝑚𝑏𝑒𝑟

𝑹𝒖𝒓𝒂𝒍 𝒂𝒏𝒏𝒖𝒂𝒍 𝒉𝒐𝒖𝒔𝒆𝒉𝒐𝒍𝒅 𝒆𝒍𝒆𝒄𝒕𝒓𝒊𝒄𝒊𝒕𝒚 𝒅𝒆𝒎𝒂𝒏𝒅

𝑁𝑜𝑛 𝑝𝑜𝑜𝑟 ℎ𝑜𝑢𝑠𝑒ℎ𝑜𝑙𝑑 𝑛𝑢𝑚𝑏𝑒𝑟×1000 + 𝑃𝑜𝑜𝑟 ℎ𝑜𝑢𝑠𝑒ℎ𝑜𝑙𝑑 𝑛𝑢𝑚𝑏𝑒𝑟×300 =  

𝑁𝑜𝑛 𝑝𝑜𝑜𝑟 ℎ𝑜𝑢𝑠𝑒ℎ𝑜𝑙𝑑 𝑛𝑢𝑚𝑏𝑒𝑟 + 𝑃𝑜𝑜𝑟 ℎ𝑜𝑢𝑠𝑒ℎ𝑜𝑙𝑑 𝑛𝑢𝑚𝑏𝑒𝑟

The result of this is a household electricity demand estimate, combining urban/rural and wealth/poverty information, for each point location. The average was also calculated nationwide (see Table 3 below), to be about 725 kWh/household per year in rural areas (which is at the low extreme of the PPL consumption data presented in Figure 8 above), and about 1,900 kWh/household per year in urban areas.^[9] Note that although Table 3 includes percentages of “poor” and “rich” homes, this is not meant to indicate that households in the country fall into only two distinct groups. Instead, the methodology of averaging demand values of percentages of “poor” and “rich” homes serves to define an average demand value for households in each community. 

2.3.2 Preparation of infrastructure demand points 

The data available for these facilities generally included little information beyond the lat/lon coordinates and facility type. The lack of additional attribute information (such as size of institution, or number of full-time staff) allowed only a very simple estimation of social infrastructure demands, as outlined below: 

                 Schools: 2,000 kWh/year 

                 Clinics / Health Posts: 4,000 kWh/year 

                 Hospitals: 40,000 kWh/year 

Another noteworthy pre-processing step taken with social infrastructure points. All social infrastructure points were assessed for their distance from “settlements” (as represented by CU points from NSO). Any education or health facility that was within 1 kilometer of an existing CU point had its electricity demand aggregated with the residential demand for that residential point. The reason for this is that the electricity needs of a smaller facility, such as a school or clinic, can often be met by the same transformer that provides power to households, so the two demands – residential and social infrastructure – can be aggregated at the location of the residential point. However, education and health facilities located more than 1 km from a CU point lie outside of the reported range (from PPL) of low voltage line connections to a transformer, and so were preserved as independent points in the dataset, distinct from residential locations. As results of modelling will later show, the vast majority of solar systems recommended by the model target these facilities.^[10]  

2.3.3 Cost and technical parameters for the grid distribution network 

Alongside household electricity demand, other critical model parameters for this analysis quantify the initial (or capex) investment costs for medium and low voltage distribution lines. In discussions with PPL on the subject of line costs, staff reported a range, high by international standards, of between 150,000 – 250,000 PGK per km of line (US$50,000 - $85,000/km). This was expressed as an “all inclusive” cost, comprising not only MV and LV lines, but also transformers, connections, and “soft costs” such as labor, transportation, taxes, design fees, etc. Translating this very general, “all inclusive” cost into specific unit costs required focused requests for both equipment component costs and “soft costs”. While numerous project documents were obtained from PPL, most of them covered technically limited projects – for example, the extension of an MV line 1 km to add a transformer – that did not include a range of cost elements. Only rarely did PPL project documents include the full range of technical features necessary for verifying all unit costs relevant for the sort of national, comprehensive, residential electrification program of the type envisioned by SE4All. Table 4 below shows one such project which included both MV and LV lines, transformers, and 200 connections with “EasyPay” meters. 

This project document was helpful for verifying cost figures that had elsewhere been given verbally, and for supplying additional cost data for technical features (such as low voltage aerial bundled cable) that had otherwise remained unclear. Costs for materials represent 4550% of the total; “soft costs” such as transportation, labor, design costs, and taxes, represent the remaining 50-55%. Additional documents such as this – with clearly enumerated and costed elements of a residential electrification program – can be very helpful in resolving cost ambiguities. In this manner, PPL project documents provided cost parameters used in the model’s calculations for grid extension. The parameter values used in the “Base Case” model scenario are presented in Table 5 below.  

system (grid)             transformer cost per grid system kilowatt       $175 per kVA                                            1

Some notes on this list:  

        Low voltage line appears to be an important and highly variable contribution to total costs: Based on current PPL data available, LV “open wire” costs are substantially higher than LV service line costs, making this an important distinction for understanding the cost build-up for residential electrification, and how LV costs vary with changing terrain throughout the country. The row “low voltage line costs per meter” in the table above represents only the “open wire” lines (LV OW), and this price of US$30/m both urban & rural areas. Other low voltage connection costs – for the service line, meter, etc. – are included in the SEL/EI model as part of the “low voltage equipment cost per connection”. PPL project documents, combined with geospatial analysis involving review of satellite imagery for inter-household distances indicate that this cost includes two main components:US$255 for the household connection (including EasyPay meter, parts, and labor) and about US$205 (ranging from $85 to $340) for service 

line. The ranges in these costs are generally due to large variations in household spacing over the national landscape. An attempt to quantify this variation is described in Assessment of Inter-Household Distance . 

        While materials costs are in the upper range of international reference values, an additional 50-55% due to soft costs places the PNG unit costs well beyond many other countries, suggesting potential to reduce overall costs. 

These costs (primarily initial costs for equipment) are summarized visually in Figure 9 below: 

Two additional factors that require attention when interpreting the model outputs for grid system costs. The first factor relates to the cost of power per kWh. It is critical to note that this cost is not the retail cost of power paid by the consumer. Instead, it is the “wholesale” or “bus-bar” cost of power, which includes the cost of generation and transmission, and can be viewed as the cost that an electric utility limited to MV and LV distribution would pay for power in an “unbundled” system with separate and independent generation and transmission systems. In this model analysis, generation is included only as a recurring cost, within the unit cost of power. This analysis uses two settings for this parameter: 

The assumption that the cost of power will be low (roughly on par with the Port Moresby and Ramu grids today) throughout the entire “mainland” area requires some explanation. Given that several systems on the PNG mainland now operate under diesel power, this assumption requires addition of high voltage transmission lines to interconnect lower cost generation (such as hydro) with areas that currently rely on diesel. This has been discussed with DPE and PPL, and found to be realistic, both because PNG has energy resources that offer lower cost generation and because a case can be made for cost-effective HV extensions of the distances required (typically 100-150 km) for many of the demand centers in PNG, once population growth is included. (The latter option, HV extensions, is discussed in more detail in the sub-section Final results for the “Base Case”.) 

The second important factor to consider is that these model results have also been modified to add a 30% medium voltage line “correction factor.” This is needed to account for the fact that MV grid lines do not follow perfect straight-line paths between communities, but rather follow more complex pathways along roads or topography. This is illustrated in Figure 10 below, where the yellow dotted line represents the model’s proposed grid line extending straight between two communities while the blue line represents the existing road which is a more likely path for grid construction. (This correction factor is discussed in more detail in 

30% Correction Factor for MV Length.) 

2.3.4 Diesel mini-grid and solar / off-grid cost and technical parameters 

The primary costs for mini-grids and off-grid / solar systems are as follows (noting that several of the components of mini-grid costs related to the local low voltage line are also used for low voltage costing for the main grid):  

Category                 Parameter 

Omits unused / null values

Parameter     Source 

(June 2016)     ^{1 PPL} 

Some notes on this list:  

        Most cost values for equipment and other inputs for mini-grids are fairly straightforward. It is, instead, issues related to system management and maintenance that tend to be highly uncertain and difficult to quantify. For systems like diesel mini-grids and solar home systems, some of the most pressing issues are not related to technical costing and design, per se, but rather the costs and difficulties related to delivering small amounts of power reliably, consistently and cost-effectively, to very distributed customers. Metering, billing and payment systems can be costly and difficult to manage. Reliability can be a problem as small systems try to serve small and highly variable loads. These 

factors hinder cost-recovery, and management in general, as PNG has experienced in the past with rural and distributed electrification programs such as the C-systems. Nonetheless, mini-grid and off-grid / solar systems are likely to play a role in PNG due to the highly distributed population, rough terrain, and other challenges to grid extension. 

2.4 The need for a geospatial planning platform 

Looking ahead to implementation of NEROP, it is crucial to recognize the needs that a multi-year, large-scale electrification roll-out plan will require for ongoing planning, tracking of implementation and assessing progress toward access goals.   

First, a key purpose of this consultancy has been to develop capacity locally for ongoing planning.  Although the national plan has been completed by this consultancy, it is strongly recommended that local planners – working within Papua New Guinea with access to local data sources – continue this quantitative, long-term, least-cost geospatial planning.  Over the 15-year timeframe of NEROP, data sources are virtually certain to change and hopefully will improve.  Much like the mapping of PPL’s distribution infrastructure helped with this planning effort, an inventory of key commercial or social infrastructure demand points – including schools, clinics, markets, and industrial sites -- could improve the specificity and detail for this plan.  Also, the national plan will need to be localized, yielding more precise results for each province, district or even LLG.  Finally, plans will need to be reassessed as more becomes known about electricity demand in households, communities and nonresidential locations.  Experience from several countries shows that the first, national-scale plan created by international consultants is only the first step in a multi-year, nationwide effort which can be improved the more it is localized and refined.   

Second, important new data capacities will need to be added within DPE and/or PPL to monitor and evaluate the implementation of the roll-out program.  Thousands of kilometers of new grid lines will need to be mapped, and their status (under construction, completed, energized, etc.) will need to be tracked.  Access rates within communities – which have been estimated in this report (see following section Estimate of current grid access) – will need to be quantified more precisely in the near term, and monitored annually over the medium to long term, to continually gauge progress toward universal access. Locations planned for grid vs. off-grid systems will need to be identified, and information will need to be shared, to avoid conflicting or duplicative project implementation and waste of funds.  Finally, DPE and PPL will benefit from careful monitoring of the costs, speed of implementation, and other factors related to a multi-billion-dollar grid and off-grid roll-out program. 

All of these needs can be met by a geospatial planning platform which is established with international assistance, but transferred to local practitioners for use and maintenance, as needed.  Such a system combines maps and tabular outputs to capture the status of electricity access throughout the country, the progress towards universal access, and location, capacity, type and cost of electricity equipment as it is planned, procured and installed.  It is estimated that such a system will require two teams working in parallel – one team of about 4 international consultants, working for 2 years to establish, train and transfer the system for local use; and another team of about 4 local workers who will over 3 years help develop the system, manage it locally, and use it to track and report on electrification progress over the span of NEROP.  The total cost for these two teams, over 2-3 years is estimated at about US$4.5 million (see Table 28, on page 98; and Table 29, page 104).   

Estimate of current grid access  

3 Estimate of current grid access 

Grid Access was estimated using a combination of PPL data, geospatial queries performed by SEL using NSO and PPL grid data, as well as other sources. 

PPL billing data provided information for the total number of residential customers (sum of domestic customers using old analog meters and new “EasyPay” meters). 

                 The average of 2014 and 2015 totals is about ~90,000 customers 

                 ~90,000 households x 5.15 people per household = ~460,000 people nationally. 

By this calculation, at least ~6.4% of the national population has a formal PPL connection which is documented within the utility’s billing records. Note that the PPL data has substantial uncertainty. For example, the data reports residential connections within each of the utility’s nearly 30 separate grid systems. The PPL data lists only 1 residential customer for the entire system of Arawa, which seems extremely unlikely. Other systems have similarly suspect data.^[11]

Other sources (WB, HIES) provide total national electricity access rate of 12-13%. If this figure is correct, it suggests that for every documented and official PPL residential account, there is another that is note documented within PPL’s records.  

SEL performed a geospatial query to determine that about 19-20% of the NSO census units fall within 1 km of the MV grid transformers. It is estimated that, for the other ~8082% of the national population, the following are true: 

                 these communities lack grid access;  

                 they are greater than 1 km from a PPL transformer; 

                   connecting them to the grid requires some addition of transformers and/or MV 

lines 

These three bolded points above are summarized in Table 7 below, which also provides a 2030 population estimate:  

Estimate of current grid access  

This estimate both establishes the magnitude of the electrification task, and suggests what some of the important challenges may be: 

4 Final Model Results  

Based on the input data described above – demand points for settlements and social infrastructure, existing MV distribution lines and equipment, and numerous cost and technical parameters – the SEL/EI team performed multiple modelling scenario runs. Most of the results described in this document are based on set “best guess” model parameters with a reduction of approximately 40% on overall unit costs for grid extension compared to PPL’s current costs. Other scenarios were run, both a “high cost” scenario representing grid costs from recent implementation (see next section, p. 46), as well as smaller changes in a sensitivity analyses (see Appendix, p. 192). While all scenarios favor grid extension as the least cost option for most of the country’s population, the geographic scope and cost of such a program varies based on assumptions. 

A few notes before more detailed look at the model results: 

4.1 Final results for the “Base Case” 

The geographic overview of the final model results for the “Base Case” scenario are shown in map in Figure 11 and Table 8 below. General features of the plan include:  

 Grid connectivity is the recommended least-cost electrification technology for most of the nation’s communities, while mini-grid is the next most important. 

        However, there are strong geographic patterns to the recommendations. Grid is dominant in the more densely populated coastal and highland areas, while minigrid systems are recommended mostly for sparsely populated areas and islands (Bougainville, New Ireland, New Britain, Western, etc.). 

Table 8New household connections by 2030, by province (Base Case) 

Table 9 and Table 10 below summarize the cost implications of the grid and mini-grid access programs on a total and per household basis. Table 9 provides a detailed breakdown of the separate parts of the electricity access program. The lowest-cost element of the program would be improved connections -- installation of EasyPay meters or new service line and connection equipment – for those households who have grid access, either as PPL customers or as “consumers” who use power but do not pay or are not properly metered or accounted for by the utility. The next group, in terms of expense, would be households within rang of a low voltage connection (within approximately 1 km of existing grid infrastructure) but who do not yet have a connection. The next two groups are comparable in terms of expense: new grid connections made through true extension of the grid (including medium voltage lines spanning distances between communities) and off-grid connections, most likely mini-grids, which may be diesel, renewable, or hybrid, depending upon the resources, demand, and other aspects that will be better determined locally, on a case by case basis. This table includes percentages of households, and costs per household and total, for each portion of the electricity access program. 

Table 9: System results and costs for 100% electricity access (grid and off-grid) by 2030 

Table 10 provides an overview with aggregate cost and technical metrics for grid and offgrid programs, omitting the improvement investments related to existing connections.  The grid extension program would require construction of around 14,500 km of new MV lines to enable around 1.5 million new connections. The initial costs for this program – including the MV and LV grid distribution network, with intensification, but excluding sub-stations – would be about US$2.4 billion, or approximately $1,600 per household, on average. These new connections would require addition of around 340 MW of new generation nationally. While the CAPEX cost of new generation is not included as an initial cost, it is included in the overall least-cost plan in the form of recurring costs – specifically, the yearly costs of power to consumers, which also includes operations, maintenance, and periodic replacement for equipment, such as transformers. A parallel mini-grid program would electrify around 570,000 households in around 8,200 communities, at a total cost of around US$660 million, or US$1,160 per connection. Again, it is crucial to keep in mind that these grid and mini-grid programs target 100% national electrification. It most likely will require an additional discussion, involving policy questions, to determine how to prioritize which 70% of the nation’s population will receive electricity access, and which 30% will not. Considering this decision on a least-cost basis, one option would be to omit roughly half of the highest cost communities from each of the grid and off-grid programs, since these cost more per connection than intensification. 

Table 10: Model result (Base Case, with 30% MV correction factor) 

Figure 12: Proposed grid extension by PPL system (Base Case).  

Table 11 below provides quantitative data for these systems, which are projected to add ~ 310 WM of new peak demand nationwide serving 1.3 million new homes with ~14,500 km of new MV lines. On a per household basis, the averages out to about 230 W per household, a quite modest amount, reflecting the relatively rural and poorer geographies served by the NEROP program. The majority of proposed new demand, household connections and MV line is recommended for the Ramu grid, which is sometimes described in terms of its separate sub-networks. 

Table 11: Households, demand, MV added to PPL grid systems (Base Case) 

These data can help to address the validity of the assumption that a single bus-bar cost can be applied throughout the PNG mainland through HV extensions. Blue rows in this table indicate mainland systems that are currently diesel-powered, and thus have relatively high recurring costs, but where the number of new consumers adds demand of at least 8-10 MW. As new, low cost generation is added to the PNG national system, the supply power provided at lower cost by hydro and perhaps natural gas will balance the added cost of the HV lines over time. In PNG, the HV extensions are likely to be ~100-200 km, for 66 kV or 132 kV lines. With demands of 8-10 MW, savings on recurring costs should recoup HV investments for many systems, particularly once demand growth is considered. 

Another output of the model is a sequenced plan for grid roll-out, which prioritizes the construction of individual MV segments based on a simple cost-benefit calculation of the amount of power demand met divided by the initial costs of line construction. Results of this prioritization, shown in Figure 13 below, indicate that MV grid extensions are recommended 

predominantly in denser areas where grid access is relatively limited – such as the highlands and some islands, such as New Britain and Bougainville. 

4.2 Costs of grid extension per decile and household (“cost build-up”) 

As stated previously, this analysis does not create a year-by-year investment program or detailed engineering design. The incremental costs presented in Table 12 below are broken down into a simple framework of 10% “deciles” that should not be interpreted as specific investments targeted for specific years, as these sorts of budgeting and implementation decisions involve other concerns – such as the availability of funds in annual budgets, and the practical capacity of PPL or private contractors to implement grid extension in a given amount of time. Instead, the information is provided to support budget planning and decision making that must consider questions such as how much grid extension to invest in (compared with other possible investments such as non-grid electrification or even other infrastructure). It describes the rising costs of grid extension as the program approaches different access targets in order to allow the PNG government, PPL as a utility, and other partners in this extension program to evaluate and prioritize the best use of scarce funds from a policy and practical perspective.  

Figure 14 below provides similar information for changing costs throughout grid roll-out, but in graphical form.

The figure emphasizes two additional factors: a) the separate cost components that contribute to the total cost per household of MV and LV grid extension (including connection costs, LV costs, and MV costs); and b) the importance of variation in MV costs specifically as part of the overall cost profile (the latter point will be explored in more detail in the annex Geospatial Plan: Sensitivity Analysis). This figure shows, in different shades of green, the very “local” costs of grid extension, including the service drop and meter, low voltage line and transformer, all investments which are made within the community being electrified. As the green curves show, there is substantial variation in these “local” costs as the program proceeds, which is due largely to the substantial differences in settlement patterns throughout the PNG landscape (discussed in more detail in Assessment of InterHousehold Distance ). In the left-most portion of the figure, for the first 10-15% of the grid extension program, connections result from intensification, and the only costs are these local, low voltage costs, so the only visible curves are green and the total costs per household are approximately US$1,000-$1,100. Moving to the right, the figure highlights the costs for MV lines that extend over the landscape between settlements. These MV costs are shown in two lines: one represents the “Base Case” assumptions, another represents the addition of the 30% “correction factor” (see 30% Correction Factor for MV Length). The figure illustrates the increasing importance of the costs of MV line per household as the grid extension program proceeds.MV line costs remain below one-third of the per household cost for roughly the first five deciles of the MV extension program, growing to 60% of more of the total cost per household in the latter two deciles.  

Table 13 below provides more detail for the “build-up” of component costs per household for grid extension. Note that half of all initial costs are “soft costs” (labor, transport, taxes, etc.). 

While the average total household cost is around $1,600-$1,650 (with the 30% correction factor), these results also present a broader range of expected per household costs, depending primarily on geographic factors, from about US$1.000 to a maximum of around $4,500.The two dominant costs on a per household basis are LV “Open Wire” line, with an average cost of around $450 per household, and MV line, with an average of around $650$700 per household, together comprising about $1,100, or nearly 70% of the total cost for a typical PNG household.  

A key focus of geographic investigations for this project has been the importance of distances between homes and the impact on low voltage line cost. This has at least two key factors. One is the difference seen in investments for connections, including the low voltage “open wire” and the service line construction. In short, the relative length of LV open wire vs. service line in any specific grid extension project is important, since their costs are very different ($40/m vs. $8-10 per m). The other is the apparently very large variation in these LV costs due to differences in household spacing throughout the country. Household spacing patterns throughout the country were examined using satellite images (described in more detail in Assessment of Inter-Household Distance ).The higher estimate seen in Table 13 above accounts for large spacing between homes visible in some provinces, particularly the highland areas.^[13]  

As described previously, this cost build-up presented above ultimately arises from cost and technical parameters which have been obtained, for the most part, from a review of PPL project documents, supplemented by discussions with PNG-based private sector project implementers, and some international comparisons. These may improve with further review and discussion, and they may change with time or with specific efforts at cost reduction, including specific procurement practices or other efforts related to implementation of a large, national extension program. Results from other scenario runs with both higher and lower grid line costs have been preserved, and these are explored in the following section. 

4.3 Results of a high cost analysis reflecting recent project costs 

This section presents an alternate scenario to the “Base Case” presented in previous sections. The following “High Cost” scenario uses the current, higher costs as seen in a selection of recent PPL-implemented electrification projects.¹³ These costs are presented as indicative of a persisting high cost situation, rather than broadly predictive of all future costs, for a few key reasons: i) the projects reviewed for this scenario have limited geographic scope (Kimbe, Popondetta and Bougainville); ii) these projects focus primarily on generation, transmission, and extension of the medium voltage grid line, what might be referred to as the grid “backbone,” and as a consequence result in relatively few household connections for each project; iii) these projects are planned, designed, procured and implemented at a much smaller scale than will be the case for a future grid expansion program at national scale. Nonetheless, the costs from these projects provide valuable insight into key costs and assumptions, and the related implications for the scope of a future grid expansion program.  

While specific costs and technical metrics for these projects differed, the overall exercise mostly confirmed the quantitative values used for the “Base Case”, noting that the “Base Case” deliberately presumed a reduction in materials, labor, and other costs in the range of 20-40%. For many components, the unit costs are largely the same for the “Base Case” and “High Cost” scenarios. Material costs of medium voltage line were somewhat higher in the latter, while costs for low voltage line were lower in the “High Cost” case. The “2X” multiplier – which assumes that total costs are generally double material costs, once nonmaterial costs such as labor, local transport, taxes, fees, design costs – was confirmed to be roughly the same for the “Base Case” and “High Cost” scenarios.  

The key differences seen in the “High Cost” scenario related primarily to the length of low voltage line per household, and costs for connection and of transformers. In general, the costs for these components of grid extension were seen to be approximately double those assumed in the “Base Case”. However, as stressed above, it is crucial to note that the projects 

connection costs are divided between high cost LV “open wire” and lower cost “service line”, and ii) inter-household distances were estimated from satellite imager review from more than 30 places throughout the country.  Another decisive factor may be that the ADB/SMEC study seems to have included initial costs of new generation in individual household costs.  As noted previously, generation costs for this SEL/EI study have been included, but as part of recurring costs, not initial capex.  

¹³ Documents from these projects, currently in progress, were obtained from and discussed with representatives of ADB and PPL with detailed knowledge of planning, procurement, and implementation.  This, allowed review of unit costs (per km of line, per household connection) at a greater level of detail than information acquired previously.  The SEL/EI team remain the source of the final conclusions regarding costs. 

reviewed were not targeting access to households, but rather construction of the grid “backbone”. 

The quantitative and map results of the “High Cost” scenario are shown in the tables and figures that follow. 

Table 14 below presents similar information to Table 10 (presented for the Base Case in the previous section).  The total number of households targeted for grid access does not change dramatically between the two scenarios: The High Cost scenario recommends 1.3 million households whereas the Base Case recommends around 1.5 million (a reduction of about 1112%. Somewhat larger differences can be seen in the reduced total extent of the grid program (~11,000 km of MV line for the High Cost scenario vs. ~14,500 km for the Base Case, a reduction of about 24%), and the average cost per household of grid connections 

($3,100 for the High Cost scenario vs. $1,600 for the Base Case, or an increase of 45-50%). 

Figure 15 below is comparable to Figure 14 (presented for the Base Case in the preceding section) and provides the perhaps surprising result that per household connection costs are likely to be more or less level beyond the 30% household connection point because as the MV line per household increases (blue and red cost curves) the LV line per household decreases (the green lines). This is due to the fact that highlands are likely to be electrified first, where population densities are high (lowering MV costs) but spacing between homes is large (raising LV costs), and other areas are likely to be electrified somewhat later, where communities are widely spaced (raising MV costs) but households are close (lowering LV costs). 

The maps presented in Figure 16 and Figure 17 show the same essential geographic patterns as were visible in Figure 11 Figure 13 (for the “Base Case” scenario), in that most expansion is seen in highland areas, on the Ramu grid, and coastal areas. However, the total geographic scope is slightly restricted in the “High Cost” case versus the “Base Case”, primarily due to the additional costs of the “last mile” of transformers, low voltage line to the home, and household connections. 

Clearly these cost differences between Base Case and High Cost scenarios are significant, particularly for the total costs per household connection, though far less so for the geographic pattern of recommended grid expansion. The more critical question from a planning perspective is whether the high costs will persist once a grid connection program is implemented at scale. There are reasons to believe that costs can be dramatically reduced, based on international experiences with sector wide grid expansion. These are discussed briefly in the following section.  

4.4 Potential approaches to cost reduction 

As noted previously based on PPL information, materials costs are high for the country, and on top of this, an additional 50% of total initial costs for grid extension arise from nonmaterial, or “soft costs,” including labor, transportation, design costs, taxes and other fees. Given the importance of high unit costs as a factor in electricity access planning for PNG, it is important to consider whether cost reductions are a possibility in the near future and, if so, what are some possible approaches to achieving these reductions. Several of these, drawn from international experience, are described in brief below. All of the following combined can lead to relatively quick and dramatic reductions -- 40% reduction in the span of 5 years is not at all unreasonable.^[14]

        Single wire earth return (SWER): Reviews of SWER costs^[15] report that SWER can bring real cost savings – of perhaps 30% or so – but that this involves tradeoffs that must be evaluated by PPL. 

This broad list represents a variety of options for ensuring that implementation at scale can be achieved at costs closer to the Base Case from this report, rather than the High Costs seen in a few recent projects.  

Part 2: Implementation of the Least-Cost National Electrification Plan  

PART 2: IMPLEMENTATION OF THE LEAST-COST NATIONAL ELECTRIFICATION PLAN

5 Institutional framework 

5.1 Introduction

In this section we review the existing institutional framework in PNG with respect to implementation of NEROP. We evaluate the existing framework and use lessons from international experience to recommend improvements to the framework. 

The institutional framework and the framework for funding the implementation of NEROP are inevitably interlinked, but in this section we focus predominately on the institutional framework and discuss the detail of the funding framework in Section 6. 

Terms of reference 

Key excerpts from the terms of reference for Task 7, the review of the institutional framework, are provided below: 

Structure of this section 

This section is structured as follows: 

                 Section 5.2 describes the existing framework in PNG. 

                    Section 5.3 describes and evaluates the main options for implementing NEROP. 

                 Section 5.4 summarises our recommendations. 

5.2 Existing framework 

5.2.1 Existing policy and regulations 

There are number of different key policy documents, some of which are due to be updated 

The existing regulatory and institutional framework for rural electrification is set out in a number of different laws, regulations, and policy documents, as summarised in the box below.  

The Electricity Industry Policy (EIP) was published in 2011 and has been due to be reviewed and updated since 2013 (it has not).  At present there is also no policy dedicated to rural electrification. 

This implies that, going forward, there is scope to revise and update the institutional framework for implementing NEROP. 

Policy sets a clear target for electrification: 70% of households by 2030 

The PNG Development Strategic Plan 2010-2030 provides clear policy direction with respect to electrification: Over 70% of households by 2030. This target is further elaborated in the Strategic Plan as 10.5% annual growth target (which means relatively few connections in the early years, but a much higher connection rate towards 2030), whereas the PNG Medium Term Development Plan 2011-2015 sets five yearly targets with a more linear connection rate. Key excerpts are provided in the box below. 

EMC is the overarching institution responsible for overseeing NEROP 

The institutional framework set out in the EIP, including the Electricity Management Committee (EMC) and the Energy Wing of DPE, has been established, although it is arguably not yet fully effective. EMC has only met a small number of times and the Energy Wing of DPE is understaffed.  

The key excerpt from the EIP is provided in the box below. 

DPE has clear responsibility for implementing NEROP 

The EIP envisaged that the Energy Wing of the Department of Energy and Petroleum (DPE) would be secretariat to the EMC and would carry out the operational functions, namely prioritising electrification projects, allocating funding, and tendering out projects (as discussed further below).  

The EIP does not envisage that DPE would actually construct and operate electrification schemes, only that it will be the implementing agency that facilitates the construction and operation contracts. 

Key excerpts from the EIP are provided in the box below. 

PPL can be directed to carry out grid extensions in non-commercially viable areas, if funding is provided 

To-date, PPL has not been explicitly obliged to carry out grid extensions, for example based on electrification targets, although the EIP does foresee this possibility so long as PPL is compensated. The EIP is not explicit about how future grid extensions will be implemented under NEROP, although it appears to assume that PPL will operate any future grid extensions (as opposed to tendering out both the construction and operation, as per minigrids). 

The EIP makes it clear that PPL will not be expected to undertake unprofitable grid extensions. PPL may apply for Community Service Obligation funding in such cases. 

PPL’s exclusive zone is only 10km of its existing network (and customers with load >10MW are excluded), so third parties are already free to apply for a license to supply power in all other areas. 

Key excerpts from the EIP are provided in the box below. 

EIP states that off-grid electrification will be competitively tendered out 

The EIP envisages that power supply to rural areas that are isolated from the main grids are tendered out, with the private sector and PPL (if it chooses) competing for capital subsidies. The tariff in those areas will be set by ICCC on a cost-recovery basis (such that, presumably, no operational subsidy is required). To-date, no new isolated/mini-grids have been implemented.  

Key excerpts from the EIP are provided in the box below. 

The private sector is to play a key role in electrification 

There is clear policy direction that NEROP will include a strong emphasis on private sector participation, from the overarching Papua New Guinea Vision 2050 down to the detail of the EIP. As above, the EIP envisages that new electrification projects will be competitively tendered.  

The private sector is already able to establish new grids outside of PPL’s exclusive area, so long as it is granted a licence by ICCC (the terms of which are not onerous). The PPP centre must approve any significant transactions involving private sector financing. The EIP foresees that the private sector will own any electricity assets it finances. 

Key excerpts from the relevant policy documents are provided in the box below. 

Provincial local governments will also play a key role in funding and possibly implementing electrification projects 

The federal nature of PNG’s Government (as defined primarily in the Organic Law) means that local governments, be it at the Provincial, District, or Local-Level Government, have some responsibilities for electricity provision.  

In particular, the District Development Authorities Act makes District Development Authorities (DDAs) responsible for providing infrastructure services, but whether or not that includes power supply is subject to a ministerial determination on a case-by-case basis. We are not aware of any such ministerial determinations with respect to power supply as yet. 

The EIP recognises that its policy directives to centralise electricity provision, either by being supplied by PPL’s grid or supply being tendered out by DPE, are potentially at odds with the Organic Law (and the resulting DDA Act) and therefore recognises that the approach may have to vary province-by-province to reflect local laws and the on-the-ground circumstances. 

Key excerpts from the relevant policy documents are provided in the box below. 

A robust licensing and tariff regulatory regime is already in place, although it may be transferred to a new entity  

The Independent Consumer Competition Commission (ICCC) has had responsibility for the economic regulation of the power sector since 2002. In our view, the regulatory regime established by the ICCC is robust and to an international standard. 

Technical regulation of PPL is less ideal. ICCC were originally delegated responsibility, but without any technical staff, ICCC delegated the role to PPL (which was a reasonable approach, given that PPL was the national supplier). The EIP envisaged third-party suppliers/retailers, which requires that technical regulation be carried out by an independent party. Responsibility was meant to be transferred to DPE, but without the Energy Wing having adequate resources this has not yet happened.  

It is important to note that in PNG political influence sometimes runs counter to regulatory rules and decisions, which impacts negatively on PPL’s performance^[16]. For example, ICCC approved tariff increases in 2013, but PPL has not implemented them (to the detriment of its financial performance), presumably due to media statements by senior politicians. Some of PPL’s recent decisions to sign up to new Power Purchase Agreements do not appear to be least-cost^[17], which may have come about due to influence by politicians and/or board members. This is certainly not an uncommon phenomenon, but underlines the importance of insulating the institutions responsible for implementing NEROP from political interference. 

Key excerpts from the EIP are provided in the box below. 

EMC and DPE will lead monitoring of NEROP, with assistance from DNPM and Kumul Consolidated Holdings 

The EIP envisages that EMC will monitor the national electrification rollout, with most of the day-to-day work being carried out by its secretariat, DPE.  

The Department of National Planning and Monitoring (DNPM) also have statutory obligations with respect to monitoring the performance of public enterprises and plays a key role as a member of the EMC.  

Treasury plays a key role in administering the Community Service Obligation (CSO) policy and fund (as described in Section 6), and would therefore also play a role in monitoring implementation of NEROP.

PPL’s shareholder, the Independent Public Business Coordination, was recently restructured and renamed Kumul Consolidated Holdings Ltd (KCH). This has implications for the governance of PPL as a whole, but does not directly affect PPL’s future role in electrification. 

Key excerpts from the relevant policy documents are provided in the box below. 

5.2.2 Existing institutional capacity 

PPL has dedicated rural electrification teams, but they have limited capacity to implement large rollout projects 

In its 2015 Power Development Plan PPL estimates that it has historically constructed power lines throughout PNG at the rate of about 70 km per annum and connected 30 new customers per km of distribution line, i.e. around 2,100 new connections per year. This is much less than what will be needed under NEROP.  

At present electrification is carried out by the Rural Electrification Services team within PPL (for Government funded projects) or by dedicated project units (for development partner funded projects).  

The Rural Electrification Services team comprises of 65 permanent staff. This includes a technical team in each of PPL’s five regions, comprising approximately six staff in each. The head office team is comprised of around 35 administrative staff. The team also makes use of more than 100 casual staff (predominantly linesmen) on a project-by-project basis. The team’s responsibilities with respect to MV/LV grid extension include investigation/feasibility studies, surveys, design, procurement and construction. Historically all of these tasks have been done internally, while in recent years PPL has started to contract out some services, namely the surveying and construction. Experience using private sector contractors is mixed. The Rural Electrification Services team does not connect households; this is done by a separate team within PPL. 

Recent experience implementing grid rollouts shows that PPL’s capacity is constrained and that delays are frequent. For example, on the Town Electrification Investment Programme (TEIP), financed through an ADB loan, PPL is used its internal staff to construct the line, yet only 3km of a total of 130km of line had been constructed in three months (as at late 2015). Some of these delays, but not all, are due to land issues. 

One of the key challenges identified by the Rural Electrification Services team is PPL’s procurement processes, which are cumbersome. For example, any project over PGK500,000 requires Board approval and the Board only meets quarterly. 

As an indication of the human resources that will be required to manage the larger scale NEROP rollout, the TIEP Project is due to implement around 2,500 new connections over two years and has a project team of 14 staff within PPL.  

The use of private sector contractors is discussed in more detail below. 

Local private sector capacity is reasonably strong 

There exists a sizeable pool of skilled local contractors, particularly with respect to rollout of the low voltage grid. The majority of these are run by ex-PPL or Western Power staff. Local contractors have been used for past rollouts, albeit not on the scale envisaged for the future.  

As a result of experiencing challenges and delays with PPL doing construction in-house, most development partner supported projects are now relying predominantly on external contractors. These include the Rural On Grid Electrification Project (ROGEP) and Tsak Valley Hydropower Project, both funded by NZAID. In both cases construction will be tendered out, with the project unit supported by external advisors. In 2015, the POM grid extension to Exxon-Mobil’s new 25MW generation facility was outsourced to private contractors and completed successfully. 

It is difficult to gauge the number of private sector contractors that are currently operating in PNG. Discussions with one such contractor revealed that there are at least a handful of contractors but that the number has been constrained by the fact that there is relatively little work for them (given the slow speed of PPL’s grid rollout). It is also noteworthy that contractors are often reluctant to work for PPL because payments are invariably delayed. Nevertheless, there is a clear willingness among these contractors to scale up operations once NEROP is implemented, but it will take time. Most will be capital constrained and staff will need to be trained. Most local contractors are currently sourcing materials from Australia and New Zealand (with poles sometimes being sourced from PNG), but as scale increases they will no doubt look towards Asia. 

The capacity of key policy-oriented institutions is currently limited 

As introduced in Section 5.2.1 above, the EMC currently has overall responsibility for NEROP and DPE, as EMC’s secretariat, plays the lead role in day-to-day implementation. Despite being established in 2011, EMC has only met a small number of times. This is arguably because of limited need, but it is nevertheless not a good indication of effectiveness. DPE may need assistance in mobilising committee members, or encouraging EMC members to delegate to less senior staff where appropriate. 

The EIP set out a clear structure for the Energy Wing of DPE, including a dedicated rural electrification division, as summarised in the figure below. However, five years on this structure has still not been established and the Energy Wing is understaffed. 

Local government capacity to implement electrification projects is low 

As discussed in Section 5.2.1, the District Development Authorities Act of 2014 makes District Development Authorities (DDAs) responsible for providing infrastructure services in provincial areas, but whether or not that includes power supply is subject to a ministerial determination on a case-by-case basis.  

DDA’s will certainly have a future role to play in funding electrification projects using local government funding (as discussed in Section 6). However, it is unlikely that they will have capacity to implement electrification projects. Technical expertise within provincial administration is low, as evidenced by the failure of almost all the 91 rural power stations that were handed over by PPL to local governments in the past. This suggests that the local government management of off-grids is likely to be problematic. We discuss the option of community participation in off-grids in Section 5.3.5 

Economic regulation is strong, technical regulation (outside of PPL) is not 

The ICCC has operated a robust, effective economic regulatory regime since 2012. Within ICCC, the Regulated Industries Division has responsibility for regulation of the power sector.  

In practice, ICCC’s main functions with respect to the power sector involve: 

                 Licensing entities 

                 Preparing regulatory contracts and carrying out the associated tariff reviews 

        Approving third-party/IPP connections to the grid, in particularly power purchase prices, connection fees, and wheeling charges. 

PPL still have responsibility for technical regulation in the industry, despite the EIP envisaging responsibility being transferred to DPE. This is largely due to the fact that there is limited technical expertise in the public sector outside of PPL. There is an urgent need to transfer technical regulation away from PPL given the increasing role that third-parties are playing the power sector. This is also important for the implementation of NEROP. 

While there are clear advantages to establishing ICCC as the dedicated energy sector regulator, in particular the economies of scope in bringing together the economic and technical regulation, there is a serious risk that ICCC’s strong capacity and track record gets diluted and that regulation becomes more politicised. This poses a serious risk to NEROP, given the importance of the regulatory framework. 

Development partners have a strong presence and capacity in the PPL sector, however there is limited coordination between the various partner 

Development partners have played a key role in funding past electrification projects. Partners active in the power sector in PNG include: 

To-date funding has occurred on a project-by-project basis and no sector-wide approach has been established, although there is significant cooperation and communication between the development partners. 

5.2.3 Summary of strengths and weaknesses 

The figure below provides an overview of the key stakeholders involved in electrification in PNG. In following table provides a summary of the strengths and weaknesses of PNG’s institutional framework, as described in the preceding sections. 

Institutional framework

Source: ECA 

5.3 Evaluation of options 

In this section we describe and evaluate a number of different institutional models that could be applied in PNG to implement NEROP.  

5.3.1 Introduction

Key ingredients of a successful model 

A number of studies^[18] have reviewed rural electrification programs internationally and identified the key reasons for their successes and failures. From these, we can identify the key ingredients of success, which are summarised as follows: 

Later in this section, we use these ingredients as a means of evaluating the institutional/ regulatory options available to PNG. 

Key characteristics of Papua New Guinea 

As above, one of the key ingredients to the successful implementation of rural electrification programs is that the institutional framework matches the specific circumstances of the country-at-hand. PNG is unique in many respects and therefore it is not realistic to simply copy the model applied in another country. Some of the unique challenges facing rural electrification in PNG are summarised in the table below and are described in more detail in Section 5.2. 

Source: ECA 

Later in this section, we use these key characteristics as a means of evaluating the institutional options available to PNG. 

Selection of key options 

The primary distinction between most countries’ approach to rural electrification is whether they took a centralised or decentralised approach. This essentially means whether the national utility played the leading role in grid rollout and the establishment of new offgrids, or whether there were multiple private or community-based operators throughout the country. 

The trade-off can be summarised as: 

“Whether the natural cost-advantages of centralised rural electrification from economies of scale (in finance, investments and management) and of scope (integration of planning, securing of investment finance and implementation) are superior to the competitive forces of unleashing multiple private actors in de-centralised RE, or vice-versa?”^[19]

The evidence from experience internationally suggests that a centralised approach is generally more successful in achieving new connections quickly, so long as the power company in charge is reasonably capable and efficient. This is not a surprising result after all, given the challenges in making rural electrification commercially viable for the private sector.  

A decentralised approach has been taken in many countries in an effort to circumvent the weak capacity of the centralised utility, with mixed results. Of the factors that contribute to the successful implementation of a decentralised approach, a clear regulatory framework and funding mechanism is the most important. The private sector will simply not be interested in investing if the framework does not minimise their risks and guarantee commercial viability. International experience shows that a decentralised approach tends to have an improved chance of success when a country is first starting its electrification program. This is because as the country gets electrified, it gets progressively more expensive to provide access and therefore more challenging to make investments viable for the private sector. 

It is, of course, possible to combine the centralised and decentralised approaches, usually by giving the central utility the key role in grid expansion and the private sector the key role in establishing off-grid solutions. In the following sub-sections, we explore four key approaches that all sit within the centralised>decentralised spectrum. 

Countries reviewed 

In identifying the key options and recommend an appropriate institutional framework for Papua New Guinea, we have reviewed the experience of numerous countries with respect to rural electrification. These include Burkina Faso, Cambodia, Chile, Costa Rica, Ethiopia, Ghana, Guatemala, Kenya, Lao PDR, Mali, Mexico, Morocco, Myanmar, Rwanda, Thailand, Uganda, and Vietnam. 

As described above, it is not appropriate just to copy one country’s approach and apply it in PNG. Therefore, in the following sections, rather than catalogue the approaches in all these countries, we focus on picking out key features that are particularly relevant to the challenges faced in PNG.  

5.3.2 Centralised implementation: PPL plays the lead role  

Overview of approach 

Under this option PPL would be the lead entity responsible for rural electrification, including undertaking grid extensions, establishing mini-grids, and installing solar home systems. PPL would be the only retail supplier nationally. In practise, this would require establishing a separate division within PPL (which is a significantly beefed up version of the existing rural electrification team), and/or possibly a subsidiary company that implements and manages the mini-grids. 

Key advantages and disadvantages 

The key advantage of this approach is that it utilises the fact that PPL has the vast majority of existing technical and commercial capacity in PNG. A fully centralised approach has the potential to deliver the quickest results by utilising PPL’s economies of scale. This is borne out in international experience, for example the success of countries such as Vietnam in the rapid electrification of the country by the national utility. Given that the private sector has never before provided retail services in PNG, it is hard to imagine that a widespread electrification plan can be implemented without significant involvement from PPL. 

However, PPL is already under significant operational and financial strain. This makes PNG in contrast to the international success stories of centralised implementation, most of which have had involved strong, well-funded central utilities. Giving PPL sole responsibility for an ambitious electrification plan under the current political and commercial environment would likely put PPL in a precarious position.  

In particular, PPL is already operating a small number of isolated systems at a significant loss, largely because political pressure prevents it from charging grid-specific tariffs (i.e. higher tariffs in remote areas).^[20] Increasing the number of these mini-grids dramatically without insulating PPL from the financial consequences would jeopardise all of its operations, unless there is a significant shift in political support for cost-recovery tariffs. The other key concern is resource capacity within PPL to undertake all of the rollout. PPL has the majority of the existing capacity within the electricity sector, but its capacity is still constrained and it would struggle to quickly establish numerous new systems in remote areas. 

Checklist evaluation 

An evaluation of a centralised approach to implementation of NEROP is summarised in the tables below, both with respect to whether it has the key ingredients required for successful implementation (from international experience) and whether it has the potential to overcome the unique challenges of PNG.  

Source: ECA 

Source: ECA 

5.3.3 Decentralised implementation: A private sector led approach 

Overview of approach 

Under this option all new off-grid solutions, including mini-grids, would be built and operated by private entities. The private sector could also potentially play a leading role extending the existing grid, whereby it is bulk supplied by PPL and awarded a separate concession to provide distribution and retail services. 

The mechanism for involving the private sector in the operation of grids (either isolated or grid extensions) would be along the lines of that already envisioned in the Electricity Industry Policy: a government entity (potentially DPE, but more likely a new entity such as an Off-Grid Electrification Authority) would tender out Build Operate Transfer (BOT) contracts to supply electricity in new areas. The private sector would compete for the contract by bidding the size of capital subsidy required to make the scheme financially viable. The contract would specify a cap on tariffs^[21] to ensure affordability²².  

Other contracting options, if BOT contracts are perceived to be too risky initially by the private sector, include: 

volume/revenue risk (for example, if tariff revenues are significant less or more than the O&M fee).  

In our view, BOT and DBO contract would be the preferred models to implement a decentralised approach. We would not recommend implementing O&M contracts, because the Off-Grid Electrification Authority (OGEA) should remain independent from operational issues, or it risks becoming a fully-fledged retailer than simply a management company. 

Key advantages and disadvantages 

The key advantages of major private sector involvement (particularly under a BOT type contract) are that it would: 

However, the reality is that PNG has no past track record with the private sector operating grids^[22] and it will likely be very difficult to incentivise the private sector to operate in remote rural areas. This is due to both the financial reality – a large subsidy will likely be needed to recover costs – and the practical difficulties of providing retail services in these areas. With time and experience (once it has been clearly demonstrated that these grids can be commercially viable under a sound regulatory framework), the private sector will be more willing to get involved, but in the early days of NEROP we expect that it will be difficult to involve the private sector. It is noteworthy that for a number of years (since the Electricity Industry Policy was formulated) the private sector has been allowed to establish private networks outside of PPL’s exclusive zone, but has declined to do so, even in areas that would be the ‘low hanging fruit’ with respect to financial viability. In our view, committing fully to a private sector led approach risks very slow results initially. This is evidenced by experience in countries such as Chile. 

Perceived risk by the private sector also means increased costs. Public utilities such as PPL will initially be better positioned to manage the risks related to providing retail services to rural areas and should therefore be cheaper. In addition, a fully private sector approach would involve significant transaction costs (due to having to tender and manage numerous contracts) and may have less access to concessionary funding (most of which lends only to the public sector). Eventually, these additional costs may be offset by competitive pressure and the efficiency of private operators, but it is doubtful that this would be the case in the early years of NEROP. 

Involving the private sector in small mini-grids would likely be easier than grid extensions. Until PPL is unbundled, it would be difficult for PPL to determine an accurate wholesale cost of power and ensure that it provides bulk supply services to the private sector on an equitable basis to its own retail business. It is noteworthy that the current regulations already allow for wheeling (i.e. third-party use of PPL’s grid to transfer power), but the private sector has not yet shown any interest in doing so, presumably because of reliability issues with PPL’s grids. Furthermore, separating out operation of new grid extensions would make cross-subsidisation more challenging and could introduce social/equity objections – customers connected to the same grid backbone and supplied by the same generation sources may end up paying significantly different tariffs. 

Checklist evaluation 

An evaluation of a decentralised approach to implementation of NEROP is summarised in the tables below, both with respect to whether it has the key ingredients required for successful implementation (from international experience) and whether it has the potential to overcome the unique challenges of PNG.  

Source: ECA 

Source: ECA 

5.3.4 Hybrid implementation: Both PPL and the private sector play important roles 

Overview of approach 

Under this option both PPL and the private sector would play a key role in implementing NEROP. PPL would continue to own and operate grid extensions, which until PPL is unbundled would be difficult to have the private sector implement. The private sector would play a key role in the construction and operation of off-grid solutions. Crucially however, if the private sector is slow to get involved initially, responsibility falls back on PPL. 

PPL would continue to carry out extensions of its main grid. New customers that are connected through grid extensions pay the same cost-recovery tariffs as other customers connected to that grid, as regulated by ICCC. The key difference from the current status quo is that PPL would be given specific rollout targets, which involve far greater numbers of new connections than are being achieved at present. This should enable greater economies of scale in purchasing of equipment and contracting out services, thereby bringing the cost of grid extensions down significantly from current levels, which are among the highest in the world. In most cases, concessionary financing would be mobilised to help with the cost (funding mechanisms are discussed in more detail in Section 6).  

As with current PPL grid extension projects that are financed by development partners, construction of the grid extensions would be contracted out to local contractors. A significant pool of local contractors does already exist and this will likely expand quickly as 

Institutional framework

demand for their services increases. However, in the initial years there may be a shortage in local contractors and some of the contracts may be large enough to attract international contractors.

Mini-grids would be opened up to the private sector, as described under the decentralised option in Section 5.3.3. The key differences under this hybrid model are that: 

As introduced in Section 5.3.3, this arrangement requires a central agency that tenders and manages all of the contracts for the off-grid solutions (with either the private sector or PPL). We propose an Off-Grid Electrification Authority (OGEA), or similar entity, as established in many countries, including Office of Rural Electrification (ORE) in Thailand and Rural Electrification Board (REB) in Bangladesh. In PNG’s case, the OGEA could be a sister company to PPL, with Kumul Consolidated Holdings as its shareholder. Another possibility would be to establish OGEA as a subsidiary of PPL, but this would create a conflict of interest if PPL were to bid against the private sector for contracts. OGEA would retain ownership of all mini-grids (and PPL’s existing mini-grids could eventually be transferred across to OGEA) and would ultimately be responsibile for service delivery. OGEA would not directly construct or operate systems, all services would be contracted out to PPL and the private sector. This ensures that the technical and operational capacity within PNG, which is a small country, is not spread too thinly. A new OGEA will still require significant resources and technical assistance, particularly in the initial years of NEROP. 

One variation on the OGEA approach is to establish separate regional OGEAs, as applied in Ethiopia where regional electricity bureaus offer technical consultation and affect expansion strategies. We do not recommend this approach because: 

        PNG is a small country and it is unlikely that there is sufficient human resource capacity to effectively operate multiple regional entities. 

        Many of the benefits of economies of scope would be lost, in particular lessons learned from experience implementing new schemes. 

Our expectation is that the first mini-grids that OGEA tenders out will be viewed cautiously by the private sector. There may not be sufficient interest in BOT-type contracts and a DBO contract may need to be considered, through market testing. Responsibility for operating those first grids may ultimately fall on PPL. But as OGEA and PPL demonstrate that a commercial, arms-length contract for operation of a mini-grid can be successful, the private sector will perceive it as less risky and become increasingly willing to compete for future contracts.  

Key advantages and disadvantages 

In our view, this hybrid model combines that key advantages of centralised implementation – economies of scale, use of existing technical and commercial expertise, the ability to directly control the speed of the rollout – with the key advantages of a decentralised model – utilisation of private sector capital and expertise, overcoming capacity constraints on the utility, reducing costs through competition, independence from political interference. Most of the key problems with applying the centralised or decentralised approaches in PNG, as described in Sections 5.3.2 and 5.3.3 above, are mitigated by hybrid model. 

There are a number of advantages to establishing a separate publicly-owned company (that has revenue raising capability) rather than having implementation administered by a Government department or regulator: 

        Foremost is that policy making decisions (including decisions about which communities are next in line for access and how funding is allocated) are kept separate from implementation responsibility, and therefore OGEA would be protected from political interference.  

 Another benefit is that an independent commercial entity such as OGEA will be better positioned to contract with the private sector, whereas Government departments are more likely to be subjected to onerous procurement rules. 

Checklist evaluation 

An evaluation of a hybrid approach to implementation of NEROP is summarised in the tables below, both with respect to whether it has the key ingredients required for successful implementation (from international experience) and whether it has the potential to overcome the unique challenges of PNG.  

Source: ECA 

Source: ECA 

5.3.5 Community involvement: Rural electricity cooperatives 

Overview of approach 

This option is a variation on the decentralised / private sector approach, and therefore can be integrated into the hybrid approach described above. Rather than relying a central OGEA owning new grids and on the private sector to operate them, communities are mobilised to form electricity cooperatives that own, and in some cases operate, the new grids. 

This option is still heavily reliant on a central OGEA (or similar) that helps mobilises communities, monitors the schemes, and provides technical support. 

Establishment of the electricity cooperative can involve the community making a financial contribution. The mobilisation of communities can be used as a key criterion for prioritising areas to electrify.  

Key advantages and disadvantages 

There are significant advantages to involving communities in electricity provision: 

specifically, the Office of Cooperative Societies was established in 2000 to “revitalize and facilitate the development of co-operative societies in the country”. 

Unfortunately, it can be very difficult to implement schemes quickly and sustainably under this model and therefore it would be risky to rely heavily on it to implement an ambitious, large-scale electrification program. 

PNG’s past experience with community-based organisations has demonstrated how challenging the model is to implement. Many co-operatives were established by the Australian Colonial Administration during the colonial era (many of which were farming co-operatives) and they all ultimately failed. That is despite them typically carrying out far less complex functions than the management of an electricity system. The reasons for the failure of past cooperatives is summarised as follows by a 2005 study^[24]: 

“Some placed the blame on the Colonial Administration for not doing enough to assist cooperatives, whilst others thought that the failure of the movement was due to over-enthusiastic bureaucratic interference in the management of co-operative affairs, which the members resented. Others theorised that the co-operative principle of open membership that coerced people of different clans or tribes (some traditional enemies) to carry on business together, was unworkable in many areas of the country. Constant factional infighting from the directors' level to the regular members, which made it impossible for most co-operatives to function efficiently, strengthened this assertion. Others still thought that the Australian business people instigated the failure of the co-operative movement in order to thwart competition. Most likely, however, the decline of the co-operative movement was due to a combination of several factors.” 

Electricity cooperatives are likely to be particularly challenging to implement in PNG, where so many small mini-grids will need to be established. Most will not be large enough to source the required managerial and technical expertise. For example, in the Philippines many of the small mini-grids that were managed by cooperatives eventually fell into disrepair and the grids were transferred back to the central utility (NPC). Many small cooperatives also suffer from politicisation. Examples can be found from many countries, for example in the Philippines board members set tariffs far too low in order to gain popularity which would benefit their political careers. This left some cooperatives in very poor financial state which often led to restructuring or the merging of two or more cooperatives. PNG’s experience with government power stations falling into disrepair once transferred to local governments is another example of the challenges of local management.  

Most of the international success stories of electricity cooperatives involve relatively few cooperatives that are quite large in size by PNG standards. For example, in Costa Rica, four cooperatives were established that led all of the rural electrification. In the Philippines, most of the cooperatives that survived are supplied by the national grid and are reasonably large distribution networks. They therefore have the necessary critical mass to establish a strong, capable management team. It would be difficult to achieve this in PNG due to its small population and the remote location of many rural communities. 

Institutional framework

Checklist evaluation 

An evaluation of an electricity cooperatives approach to implementation of NEROP is summarised in the tables below, both with respect to whether it has the key ingredients required for successful implementation (from international experience) and whether it has the potential to overcome the unique challenges of PNG. 

Source: ECA 

Source: ECA 

5.3.6 Summary of options 

A summary of the main institutional framework options for implementing NEROP are summarised in the figure below. 

5.4 Summary of recommendations 

5.4.1 Overview of recommended framework 

We recommend that PNG adopt a hybrid approach to implementing NEROP, with PPL responsible for grid extensions (centralised implementation) and the private sector responsible for establishing new off-grid solutions, including mini-grids, (decentralised implementation). If there is insufficient interest from the private sector in establishing minigrids, the obligation falls back on PPL. 

As detailed in Section 5.3 above, this approach combines the key advantages of centralised implementation – economies of scale, use of existing technical and commercial expertise, the ability to directly control the speed of the rollout – with the key advantages of a decentralised model – utilisation of private sector capital and expertise, overcoming capacity constraints on the utility, reducing costs through competition, independence from political interference.  

Our recommended institutional structure is summarised in the figures below. 

Source: ECA 

Our specific recommendations are provided below. 

5.4.2 Implementation of grid extensions 

PPL should be responsible for grid extensions, with defined targets and funding assistance 

We recommend that PPL continue to be responsible for extending its grid to service new areas and for operating those grid extensions. Our proposed changes to the status quo include: 

PPL would also be responsible for establishing any new large grids under NEROP (perhaps defined as those requiring medium or high voltage lines), on the basis that these grids will require a higher level of technical expertise and should largely be commercially viable without operating subsidies. In accordance with the EIP, third-parties are still free to establish their own grids outside of PPL’s exclusive zone at any time. 

Institutional framework

PPL will in theory also be eligible for operational subsidies for its grids if needed for affordability, as per the mechanism described below for mini-grids. We discuss affordability in more detail in Section 6. 

PPL will need to rely heavily on the private sector contractors to carry out the grid extensions, given constraints on the capacity of PPL’s internal technical teams and procurement rules imposed on PPL. 

PPL should receive significant technical assistance, in particular to boost its Rural Electrification team’s ability to procure, manage, and monitor private sector led grid extensions.  

If the ability of PPL to competently procure and manage private sector led grid extensions continued to be a serious concern going forward, a fall back option could be for the Off-Grid Electrification Authority (described below) to procure and manage construction of all grid extensions (in addition to mini-grids), and then turn over their ownership and operation to PPL. We do not recommend OGEA leading all procurement as the preferred approach because, in our view, it is more important to focus on improving PPL’s procurement capacity than finding ways to circumvent it. If PPL is unable to procure low and medium voltage grid extensions effectively or efficiently, then there seems little hope that it could implement new generating capacity and high voltage transmission line projects, both of which are also vital to the success of a large scale rural electrification program.  

5.4.3 Implementation of off-grid solutions 

What is the Off-Grid Electrification Authority and what is its main role in NEROP? 

We recommend that the private sector be responsible for establishing new off-grid solutions via a new Off-Grid Electrification Authority (OGEA). We propose that OGEA be established as a sister company to PPL, with Kumul Consolidated Holdings as its shareholder. OGEA will be the owner and licensed operator of all off-grid solutions. OGEA will not directly construct or operate systems, all services would be contracted out to either the private sector or PPL. It will effectively be a procurement authority/asset manager. This ensures that the limited technical and operational capacity within PNG is not spread too thinly. 

There are numerous international examples of dedicated rural electrification agencies, including Philippines, Nepal, Thailand, Ethiopia, Kenya, Tanzania, Uganda, Mozambique, Guinea, Mali and Burkina Faso^[25]. Of these examples, Philippines and Mali are probably the best example of a rural electrification agency contracting services out to the private sector.  

Only some of these international examples (including Ethiopia, Mozambique, Guinea, and Nepal) have a dedicated off-grid agency, due to the fact that off-grid plays a much smaller role in most countries than in PNG. PNG is reasonably unique in that off-grid is likely to play a more prominent role (approximately 30% of households) than in other countries. This is due to a combination of reasons, including the geography of PNG, the density and distribution of the population, and the costs of supply. This means that the OGEA will be facilitating the implementation of a whole range of off-grid electrification options (including off-the-shelf solar products, solar home systems, and mini-grids). 

What are OGEA’s responsibilities? How does it interface with other institutions? 

The primary responsibilities of the OGEA will include: 

The specific responsibilities with respect to funding and procurement, alongside the responsibilities of the other key institutions, are summarised in the figures below. 

Source: ECA 

Figure 24 Overview of funding responsibilities 

Community involvement will be an important ingredient to successfully operating small, remote grids. While we do not recommend relying solely on cooperatives to manage and operate grids, we do recommend that one of the OGEA’s responsibilities includes upskilling and resourcing the local private sector, for example landowner companies. 

What off-grid solutions will the OGEA implement? 

One of the responsibilities of the OGEA will be to assess the different off-grid solutions that are available and identify the appropriate solution for each community at the project preparation stage. The tiers of electricity access and the range of technologies that can be used to deliver access, all of which are to be considered by the OGEA, are summarised in the figure below.  

Source: ESMAP, SE4ALL, 2015. Multi-tier Matrix for Access to Household Electricity Supply  

Are there alternatives to the OGEA model? Why do we recommend OGEA? 

As above, we propose that the new OGEA be established as a sister company to PPL, with 

Kumul Consolidated Holdings as its shareholder. This structure keeps OGEA separate from PPL, which is responsible for grid extensions (through a separate division or perhaps subsidiary). 

The main advantages of this approach are that: 

The main disadvantage of this approach is that it requires establishing a whole new entity. 

A number of alternative institutional options exist, as summarised in the table below.  

Table 26 Alternative institutional models to OGEA 

Source: ECA 

On balance, we consider that the OGEA model is best suited to PNG because it combines the key advantages of PPL led implementation – economies of scale, use of existing technical and commercial expertise, the ability to directly control the speed of the rollout – with the key advantages of a private sector involvement – utilisation of private sector capital and expertise, overcoming capacity constraints on the utility, reducing costs through competition, independence from political interference. 

How will OGEA contract out construction and operation of off-grid solutions? 

OGEA will, under the guidance of lead policy making agency (DPE), determine which offgrid solution (mini-grids, solar home systems etc.) is appropriate for each community that is to be electrified. The approach that OGEA will take to contracting will differ depending on the solution.  

Household level solutions, for example solar home systems, will be relatively straight forward and can be tendered out through O&M contracts / dealer models. There are numerous international examples that experience can be drawn from, for example Bangladesh, Sri Lanka, and the Philippines. Mini-grid solutions are more complex and therefore require more complex contracting arrangements. 

We expect that OGEA will tender out construction and operation of new mini-grids as either Build Operate Transfer (BOT) or Design Build Operate (DBO) contracts. Interested parties will compete for the contract by either by bidding the size of capital subsidy required to make the scheme financially viable (under a BOT) or the lease-fee it is willing to pay for the rights to operate the system (under a DBO).  These are described in more detail below: 

PPL will be allowed to compete with the private sector for OGEA tendered contracts, on the same terms. In the event that there is insufficient interest in a contract, the obligation would fall back on PPL as the operator of last resort under a DBO contract. Importantly, PPL would fully recover all of its costs under the contract. In other words, PPL will not be forced to cross-subsidise loss-making off-grids.  

How might the OGEA be structured? 

The organisation of the OGEA should be a matter of further study, but international examples provide an indication of possible suitable structures. One such example, which has similar responsibilities to the proposed OGEA, is the Rural Electrification Agency in Tanzania. Its structure is summarised in the figure below. One possible amendment to this for the OGEA would be to remove the Director of Policy and Research, given that these responsibilities would fall on DPE.  

Figure 25 Organisational structure – REA in Tanzania 

5.4.4 Policy and oversight 

The policy-making agency, DPE, will oversee NEROP 

The lead policy-making agency that will be responsible for NEROP is DPE.  

DPE’s primary role with respect to NEROP will be: 

and PPL), rather than DPE, will have primary responsibility for ensuring that procurement and safeguard requirements are met.  

        Encourage and mobilise the private sector, including involving communities where possible. 

DPE will need to work closely with a number of other government departments, in particular with: 

It is crucial that DPE is better resourced than it is currently.  

5.4.5 Technical assistance and capacity-building 

In this section, we set out an initial assessment and estimated budget for capacity-building and other technical assistance associated with the implementation of NEROP over its initial three years from 2017 to 2019. There is likely to be a continued need for support of various kinds after this initial period, but at a reduced level. This continued support will also need to respond to the lessons learned from implementation of NEROP over the initial period which, at this time, we cannot anticipate. 

Assessment of requirements 

While many of the activities required to initially implement NEROP are not necessarily new to PNG in themselves, existing capacity to meet NEROP’s requirements is weak. This is partly due to a history of understaffing and inadequate financial resources and partly due to the much greater scale of activities involved in implementing NEROP meaning a risk of overwhelming existing resources. In addition, NEROP will involve the creation of a new agency, OGEA, to manage off-grid electrification and whose capacity will largely need to be built from scratch. 

Below, in Table 27, we summarise the key institutions and entities involved in implementing NEROP, their core functions and identify whether these represent a continuation or scalingup of existing responsibilities or new functions for which greater support will be required. In developing this summary, we have assumed that planning responsibilities are split across DPE, PPL and OGEA as set out in Box 1. 

We envisage that much of NEROP will be implemented through the private sector, either acting as contractors to PPL for grid extension or as developers and operators of off-grid systems contracted to OGEA. We have not identified specific training or other technical assistance for the private sector to undertake these roles as we assume that this will be organised and provided by the responsible public sector institutions (DPE, PPL and OGEA). 

Monitor and report on implementation of NEROP by PPL, OGEA     Existing responsibility but 

and LLGs                                                                                       scaled-up^c

Detailed geospatial grid extension planning                                      Existing responsibility – training 

provided under current project 

Implementation of grid extension (directly or via private Existing responsibility contractors) 

Operation of rural electricity grid                                                     Existing responsibility