Pacific power and water utilities: the fastest returns are not in the reform programme
Across the power utilities reporting to the Pacific Power Association, system losses in the 2023 financial year ranged from 1.5 percent to 40.4 percent. Thirty-nine service areas, the same ocean, broadly the same technology, and a spread of almost forty percentage points. That one statistic frames the Pacific utility question better than any amount of discussion about scale.
The standard diagnosis of state-owned utilities in small Pacific nations is structural, and the structural case is a strong one. Customer bases of tens of thousands, spread across atolls separated by hundreds of kilometres of ocean. Diesel supplying more than half of generation in most systems and over ninety percent in some. A recent assessment from UNSW's Institute for Climate Risk and Response puts Pacific spending on fossil fuel imports at 10 to 25 percent of GDP, against regional averages of roughly 8 percent on education, with about USD 1 billion a year going to diesel for electricity alone. Then add the disaster cycle: Cyclone Pam cost Vanuatu around USD 450 million in 2015, roughly 64 percent of GDP, and catastrophe insurance covered less than one percent of the damage.
All of that is true, and none of it is going away. But structure explains the level of Pacific utility costs. It does not explain the spread. Utilities operating under near-identical constraints are producing results that differ by a factor of twenty on the single metric that most directly converts fuel into revenue. That gap is not geological. It is managerial, and a meaningful part of it is addressable without new legislation, new tariffs, or a new institutional architecture.
The portfolio numbers are worse than the anecdotes
ADB's Finding Balance 2023, the most recent edition of the benchmark study of Pacific state-owned enterprise portfolios, found that only two of nine country portfolios earned a return covering their cost of capital over 2015 to 2020. Three recorded average returns on assets or equity below zero. These are not marginal businesses making a thin margin; a large part of the Pacific SOE asset base is destroying value on a sustained basis, and utilities sit at the centre of it.
The fiscal consequences are concentrated and large. Palau Public Utilities Corporation posted an operating loss of USD 6.87 million in FY2019, equivalent to 2.6 percent of GDP, with cumulative five-year losses of around USD 21 million, or roughly 8 percent of 2019 GDP. The proximate cause is documented and uncomfortable: Palau's parliament prohibited electricity and water tariff adjustments from 2017. At the other end of the size range, ABC News reported in August 2025 that PNG Power carried liabilities of around USD 1.5 billion, alongside allegations of trading while insolvent.
Water is in a similar position with worse instrumentation. The Water Authority of Fiji reported non-revenue water of about 47 percent in 2023, easing to roughly 45 percent in the first half of 2024, against the Pacific Water and Wastewater Association's own benchmark of 25 percent. Solomon Water sits near 50 percent. Solomon Water also increased connections by 86 percent between 2019 and early 2024, which is a genuine achievement, and ADB's own project reporting still describes non-revenue water as the utility's greatest challenge. Access and efficiency are not the same programme, and the Pacific has been much better at the first.
The reform agenda is right, and it is slow
Cost-reflective tariffs, independent and competent boards, properly costed community service obligations, and separation of ownership from policy: this agenda is correct, and ADB and its partners have been advocating it consistently for over a decade. It is also multi-year work that spends political capital in small polities where the utility chairman, the minister and the largest customer are frequently within two degrees of each other.
It is worth adding that the textbook version of the agenda can be wrong here. PNG's National Research Institute, in a January 2025 review of electricity reform in small island developing states, explicitly declines to recommend the 1990s full-unbundling model for small isolated systems, on the grounds that replicated corporate overheads swamp the limited competition benefit available in a market of a few hundred megawatts. In systems this small, structural reform is not automatically the high-return option. That makes the question of what to do in the meantime more important, not less.
Five things that could move inside eighteen months
1. Chase commercial losses before capital losses. The single best-documented Pacific turnaround of the last two years involved almost no capital expenditure. Marshalls Energy Company reports cutting system losses from 26.5 percent to 19.5 percent in a year, recovering close to 9.6 million kWh and back-billing around USD 4.1 million, driven by more than 1,300 field inspections between June 2024 and January 2025 and a new revenue protection function placed inside internal audit. Whatever the final audited number, the shape of the intervention is instructive: inspections, meter integrity, anomaly detection and back-billing are cheap, fast, and repeatable. Development partners fund very little of this, because it looks like operating expenditure rather than a project.
2. Fund the meter-to-cash chain, not just the asset. Fiji attributed its two-point non-revenue water reduction to leak detection, billing accuracy and customer engagement rather than to new pipe, and in September 2024 awarded a performance-based water loss contract for the Suva-Nausori corridor. Tonga Power has been rolling out prepaid smart meters since 2016. Both are the right instinct. What is missing is a standing facility that will finance metering, billing systems, collections and revenue protection at the same scale and with the same patience as a generation asset, and that pays on measured loss reduction rather than on units installed.
3. Rebuild the regional information base. The Pacific Power Association's FY2023 benchmarking report reached the public in October 2025. On the water side, there is no equally accessible recent regional benchmarking edition at all; the data now flows largely into the World Bank's IB-NET rather than into a visible annual publication. A two to three year reporting lag makes benchmarking useless as a management tool and nearly useless as an accountability tool. Restoring a current, published, comparable dataset across Pacific power and water utilities is a small line item with an unusually high leverage ratio, because everything else on this list depends on being able to measure it.
4. Pool the scarce functions, not the assets. The binding constraint in most Pacific utilities is not capital, it is people with specific skills, and there are not enough of them anywhere in the region to staff every utility separately. Two vehicles already exist to work with: the Office of the Pacific Energy Regulators Alliance, which shares regulatory capacity across jurisdictions, and the Pacific Partnership for Energy Security proposed with the Pacific Power Association in September 2025, which includes a regional project management unit supplementation scheme. That last element deserves more attention than it has received. The Lowy Institute has documented a Pacific planning office of seven staff in which a single official managed relationships with 58 aid partners. Pooled procurement of solar modules, batteries and grid equipment belongs in the same category.
5. Automate the fuel pass-through and put a price on the community service obligation. Where an enabling tariff order already exists, moving fuel cost adjustment onto an automatic formula is administrative rather than legislative, and it removes the single most common trigger for the tariff freezes that bankrupted PPUC. Palau has since moved from an automated fuel price adjustment clause to a distributed energy rate. Separately, ADB's own work finds that existing community service obligation arrangements across the region give weak incentives to serve high-cost areas while diluting commercial focus. Samoa's framework, in which obligations are costed including a reasonable margin, contracted and separately funded, is the regional model and could be transplanted faster than it has been.
The caveat, and the point
None of this list is a substitute for reform, and it would be easy to oversell. The Marshall Islands result is self-reported by the utility and a separate early-2025 source still cites losses above 22 percent. Utility twinning partnerships, which pair Tonga Water Board with Unitywater and Solomon Water with Goulburn Valley Water among others, are well documented as arrangements and, as far as I can establish, have never been subjected to a published quantified outcome evaluation. That absence is itself telling about how the sector measures its own interventions. And PNG Power is the standing demonstration that operational fixes cannot outrun a governance failure indefinitely.
The point is one of sequencing. Tariff reform, board reform and CSO contracting are the things that make a Pacific utility solvent over a decade. Loss recovery, billing integrity, current benchmarking and pooled scarce capacity are the things that buy the decade. At present the development finance architecture is heavily weighted towards the first category and towards capital assets, which are visible, financeable and attributable, while the second category falls between operating expenditure and technical assistance and gets funded thinly by everyone.
For ADB and the other Pacific partners, the practical test is straightforward: how much of the current portfolio pays for revenue recovered rather than assets installed. On present evidence the answer is very little, and that is the cheapest thing on this list to change.
Figures are drawn from published Pacific Power Association, ADB, PRIF, Lowy Institute and utility sources; the interpretation is my own.
