The Entropy of Energy Regulation in the Pacific

The Pacific energy debate is usually framed as a problem of generation technology: diesel versus solar, batteries versus thermal plant, utility-scale projects versus distributed energy. Those choices matter. But beneath them sits a less visible system: the machinery that converts evidence into licences, tariffs, standards, procurement decisions, investment approvals and public accountability.

That machinery can lose coherence.

A fuel-price shock is absorbed temporarily rather than passed through. A donor-funded solar project is procured under one set of assumptions while the tariff model is maintained under another. A battery is treated as generation in one decision, network infrastructure in another and an unclassified asset in the accounting system. An emergency subsidy begins as a temporary response, then becomes politically difficult to remove. A consultant’s model is delivered, but the data dictionary, assumptions and decision history are not inherited by the next team.

No single decision necessarily looks irrational. Yet, taken together, the system becomes harder to understand, predict and correct. This is the regulatory equivalent of entropy.

1. Entropy is a governance metaphor, not a thermodynamic claim

In thermodynamics, entropy is associated with the dispersal of energy and the declining availability of that energy to perform useful work. In information theory, entropy represents uncertainty. Applied carefully to regulation, the metaphor describes the loss of decision-useful order.

Regulatory entropy is not simply the number of rules. A country may have many rules and still have a coherent system. Nor is institutional diversity automatically a weakness. The problem arises when mandates, datasets, exceptions, contracts and temporary measures multiply without a common architecture.

A lower-entropy regime has four characteristics: the responsible decision-maker is identifiable; the data are current and reproducible; the allocation of costs and risks is explicit; and the decision can be traced, reviewed and updated without rebuilding the system from scratch.

A higher-entropy regime produces several plausible versions of the truth. Different institutions may hold different demand forecasts, asset registers or fuel assumptions. Legal authority may be distributed, but the sequence for exercising that authority may be unclear. Decisions are delayed, and delay itself begins to determine outcomes.

The Regulatory Entropy Index is proposed here as a diagnostic concept, not as an established published metric. Its purpose would not be to reward institutional centralization. Shared authority with explicit sequencing, interoperable data and disciplined deadlines could score better than formally centralized authority exercised through opaque processes.

2. Why entropy rises quickly in Pacific energy systems

Scale does not mean simplicity

Small island power systems have fewer customers over which to spread the fixed costs of regulation, engineering, cybersecurity, forecasting, procurement and legal review. The World Bank notes that around half of the small-island systems in its sample had installed capacity below 60 MW, while remoteness, limited resources, climate vulnerability and the small size of projects raise financing and supply-chain barriers.¹

A 20 MW system does not need one hundredth of the regulatory architecture of a 2,000 MW system. It still requires a grid code, safety standards, tariff methodology, investment tests, procurement discipline, consumer protection, metering rules, disaster planning and competent contract management. Some functions become more demanding because there may be no deep local market of engineers, economists, lawyers and system operators from which to draw.

Fuel shocks turn the regulatory calendar into an emergency calendar

Imported fuel remains a dominant cost driver in many Pacific utilities. In the Pacific Power Association’s FY2023 benchmarking data, fuel expenditure was the largest cost component for most reporting utilities and had a median share of 69.7% among the 16 utilities for which the measure was available.²

This creates a recurring institutional dilemma. A rapid pass-through protects utility cash flow but exposes households and businesses to volatility. A delayed pass-through protects customers temporarily but creates under-recovery, working-capital pressure and eventually a larger adjustment. A subsidy protects both customer and utility only by transferring the obligation to the public budget.

Regulation cannot abolish cost. It can only allocate cost across customers, taxpayers, creditors, donors, future consumers and service quality.

Every tariff freeze therefore has an accounting destination. Where that destination is not stated, entropy increases because the apparent price and the economic cost diverge.

Climate resilience sits awkwardly inside conventional regulation

Cyclones, flooding, salt-air corrosion, landslides and maritime logistics affect the prudent design and lifecycle cost of Pacific electricity assets. Yet traditional tariff and procurement systems often reward the least-cost compliant option at the point of purchase, not the least-regret option over a climate-exposed asset life.

The World Bank argues that resilience should be incorporated throughout project development and that planning should move from a narrow least-cost approach toward least-regret planning. Its review found that only about 15% of selected SIDS had begun incorporating resilience into energy planning.³

When resilience is not recognized explicitly, it reappears elsewhere: in emergency appropriations, donor appeals, unplanned outages, insurance exclusions or shortened asset lives. This is another entropy transfer—the cost remains, but its timing, owner and regulatory treatment become uncertain.

Institutional diversity creates interfaces

The Pacific does not have a single regulatory model. Among the 18 utilities reporting to the PPA for FY2023, 11 were classified as self-regulated public utilities, five as externally regulated public utilities, and two as externally regulated private or concession utilities.⁴

This diversity is rational. Island systems differ in legal history, population, ownership, utility scale and administrative capacity. But every model generates different interfaces among government, utility, regulator, customers, lenders and development partners.

Independence should therefore be understood as more than the formal location of an office. In practice, regulatory independence is a method: publish the data, disclose the assumptions, explain the allocation of risk, record reasons for the decision and make the next review predictable.

The energy transition adds actors faster than institutions adapt

A diesel-dominated vertically integrated utility can be governed through a relatively narrow set of relationships. A modern transition adds independent power producers, rooftop solar, batteries, electric vehicles, charging operators, aggregators, smart meters, data platforms and potentially time-varying tariffs.

Each new technology creates classification questions. Is a battery generation, network support or a competitive service? Who pays for grid reinforcement triggered by public charging? How should curtailment risk be allocated in a small grid? Who certifies imported chargers, reused batteries or inverter settings? Which institution owns operational data produced by a donor-funded platform?

The transition can lower energy costs and fuel exposure while simultaneously increasing institutional entropy. Technology is not the problem. Unintegrated interfaces are.

3. What regulatory entropy looks like in practice

Regulatory entropy is visible through recurring symptoms: tariff decisions repeatedly deferred and replaced by temporary adjustments; multiple asset registers or demand forecasts with no authoritative version; subsidies whose fiscal cost, beneficiary and expiry rule are unclear; power-purchase contracts negotiated without a published integration framework; standards copied from larger markets without a Pacific implementation pathway; and data requests repeated by each consultant because prior datasets and definitions were not inherited.

PPA’s FY2023 benchmarking report illustrates the size of the challenge. It found large differences in reported residential bills, system losses ranging from roughly 1.5% to above 40%, and revenue below estimated cost of service for almost all surveyed utilities. It also reported an association between stronger governance practices and better financial and operational performance, while caution is needed before treating that association as causal.⁵

For a residential customer consuming 500 kWh per month, the reported bills ranged from US$75.71 to US$425.06—a ratio of approximately 5.6 to one.⁶

The important question is not why every tariff is different. They should differ where underlying costs and public choices differ. The more revealing question is whether the reasons for those differences are legible.

A low tariff may reflect hydro resources, efficient operations or an explicit social policy. It may also reflect deferred maintenance, unrecovered fuel cost or an unrecorded government obligation. A high tariff may indicate inefficiency, but it may also reflect honest cost recovery in a small remote system. Price alone cannot distinguish these stories. Transparent regulatory data can.

4. A practical entropy diagnostic

A Pacific government, utility or regulator could assess entropy without creating a complicated new bureaucracy. A yearly diagnostic could score the following dimensions from zero to four:

The output should be a management instrument, not a ranking exercise. The aim is to locate where uncertainty is accumulating and where a relatively small intervention—such as a common data dictionary or an automatic fuel true-up—could materially improve the whole system.

5. Seven design principles for lower-entropy regulation

1. Build one regulatory data spine

Utilities, ministries, regulators and project teams should not maintain separate definitions of generation, sales, losses, customer classes, fuel use and asset condition. A common data dictionary, controlled source files, version history and decision-level dataset should sit beneath tariff reviews, investment plans and transition modelling.

The World Bank identifies inadequate management, regulation, planning tools and data capability as constraints on renewable integration in SIDS.¹ The solution is not merely more data collection. It is the creation of a data lineage: each published number should be traceable to a source, owner, definition, date and transformation.

2. Separate cost recovery from social protection

Affordability is a legitimate public objective, but hiding it inside an under-recovering tariff weakens both social policy and utility finance. The efficient cost of service should be calculated transparently. Any social protection should then be delivered through an explicit lifeline block, targeted transfer, community-service obligation or budgeted subsidy.

This makes the policy choice visible. It also prevents the utility balance sheet from becoming an unofficial welfare budget. Fuel and foreign-exchange adjustments should use pre-agreed formulas, thresholds, review intervals and true-up mechanisms. Predictability lowers both political shock and financing risk.

3. Regulate resilience as a service, not as an afterthought

Resilience should have a defined regulatory treatment. Standards can specify corrosion resistance, wind loading, ingress protection, backup communications, black-start capability and emergency disconnection. Investment tests can assign value to avoided outage duration, faster restoration and reduced dependence on vulnerable fuel logistics.

A resilience allowance or approved asset category should be conditional on measurable outputs. This prevents “resilience” from becoming either an unfunded mandate or an unexamined premium.

4. Pool scarce expertise regionally while retaining national authority

Not every Pacific jurisdiction needs a permanent in-house specialist for every issue. Regional panels could provide shared tariff modelling, grid-code review, power-purchase agreement analysis, battery safety, cybersecurity and appeals support. National institutions would retain legal authority while drawing on a repeatable common service.

The Office of Pacific Energy Regulators Alliance already provides a regional platform intended to strengthen regulatory decision-making and technical capacity.⁷ The next step is to convert regional cooperation into durable shared technical infrastructure: model clauses, standard data templates, benchmark methods and standing expert rosters.

5. Add a regulatory-integration clause to every major project

Every donor-financed or privately financed energy project should state, before approval, how the asset will be classified and treated in tariffs; which technical and data standards it will follow; who owns and maintains operational data; how lifecycle and decommissioning costs will be funded; what local capability and documentation will remain; and how the project interacts with existing licences, grid codes and procurement rules.

The World Bank similarly recommends transaction support that combines robust planning and technical standards with the retention of local capacity beyond the advisory engagement.⁸

6. Make exceptions expire

Emergency measures are sometimes unavoidable. Entropy arises when they become permanent through inertia. Fuel subsidies, tariff freezes, waivers, pilot licences and temporary procurement rules should include a sunset date, measurable exit conditions and a method for reconciling accumulated balances.

An exception without an expiry rule is not temporary policy. It is hidden institutional redesign.

7. Publish a Regulatory Entropy Statement

Alongside an annual tariff or sector report, the responsible institution could publish a concise statement showing the current mandate map; age and completeness of critical datasets; outstanding tariff or licence decisions; cumulative under- or over-recoveries; active subsidies, waivers and guarantees; material departures from standard methods; climate-resilience gaps; and actions taken to retire obsolete rules and datasets.

This would make institutional complexity observable. What is observable can be managed.

6. The deeper implication: design for reversible complexity

The goal should not be a perfectly static regulatory system. Such a system would be incapable of absorbing new technology, shocks or social priorities. The objective is reversible complexity: the ability to add a battery rule, an EV tariff, a new procurement model or a disaster measure without losing the capacity to explain, audit, revise or remove it.

That requires modular rules, common data, explicit interfaces and institutional memory.

The Pacific energy transition will not be governed successfully by importing the institutional bulk of large markets. Nor will it be governed by assuming that small systems need only light regulation. The more useful principle is proportional capability: the architecture must be compact, but the critical functions must be complete.

The greatest risk is not that the Pacific lacks energy plans. It is that the distance between plans, prices, physical assets and accountable decisions continues to widen.

Regulatory entropy gives that distance a name.

And once named, it can be designed against.