Most governments have historically kept their books the way a household keeps a chequebook: money in, money out, balance at year end. Cash accounting is simple, verifiable and hard to manipulate at the level of individual transactions. Yet it is systematically blind to the events that matter most for fiscal sustainability. These include promises made today but paid decades later, capital consumed quietly as roads and pipes age, and revenues earned but not yet collected.

Accrual accounting attempts to close that gap by recognising transactions when the underlying economic event occurs rather than when cash changes hands. Since New Zealand's reforms around 1990, a substantial group of governments, including Australia, the United Kingdom, Canada and France, have adopted accrual financial statements. The International Public Sector Accounting Standards have codified accrual principles for public entities. The movement is usually framed as modernisation, with governments finally adopting the discipline of the private sector.

That framing is incomplete. Accrual information is plainly richer than cash, so the useful questions lie elsewhere. Which economic realities should it capture in a sovereign context? How does it alter the shadow prices facing budget decision-makers? What does it cost to produce reliably? A government is not a firm. It taxes rather than sells, owns assets that have no market, and carries obligations defined by legislation rather than contract. Designing accrual systems well requires taking those differences seriously.

Recognition Principles: Translating Accrual Concepts to the Sovereign

The central accrual principle is deceptively simple. A transaction is recognised when the economic event occurs, not when cash moves. In government, applying this principle exposes three categories of difficulty that have no clean analogue in the firm. The first is obligations stretching across generations. The second is assets that will never be sold. The third is revenues that arise from sovereign compulsion rather than voluntary exchange. Each forces standard-setters to make judgements that are partly technical and partly constitutional.

Consider long-term obligations. A defined benefit pension promise to a civil servant accrues as service is rendered, yet a cash system records nothing until retirement. Accrual recognises the present value of benefits earned to date. The same logic extends to nuclear decommissioning, environmental remediation, loan guarantees and availability payments under public-private partnerships. The harder question is which promises count. Standard-setters treat employee pensions as exchange-based obligations. Social security, by contrast, is generally recognised only once eligibility for the next payment has been met, because legislatures can revise it.

Infrastructure presents the mirror-image problem. Roads, bridges and water networks are capitalised and depreciated, which turns the consumption of capital into a recorded expense. This matters. A government that defers maintenance can flatter its cash position for years, while an accrual statement reveals the erosion of service potential. Yet these assets have no market price and generate no separable cash flows. Valuation typically relies on depreciated replacement cost, which is essentially an engineering estimate of what it would cost to rebuild the remaining service capacity.

Tax receivables raise distinctive issues because taxes are non-exchange revenue. Under accrual principles, revenue is recognised when the taxable event occurs and the amount can be measured reliably. For income tax, the event is the earning of income. For consumption taxes, it is the underlying sale. Because assessment and collection lag these events, governments must estimate accrued revenue net of expected non-collection. This makes revenue more responsive to the economic cycle. It also introduces estimation error, particularly for self-assessed taxes with long filing lags.

The unifying insight is that accrual recognition in government is not a mechanical transfer of corporate rules. Each category embeds a judgement about which commitments are binding, which assets carry value and when the state's claim on private income becomes real. Notably, the most consequential public asset of all, the power to tax future economic activity, appears on no balance sheet. The boundary of public sector financial statements is therefore a deliberate design choice, and it shapes what policymakers can see.

Takeaway

The boundary of a government balance sheet is chosen rather than discovered. Every recognition rule quietly encodes a view about which promises the state considers binding.

Decision Relevance: How Accounting Sets the Shadow Price of Policy

Cash budgets create a predictable incentive structure. The cost of a policy is booked when paid, so any action that shifts cash beyond the budget window appears costless. Timothy Irwin's catalogue of fiscal stratagems makes the pattern clear. It includes asset sales booked as revenue, pension enhancements granted in lieu of wages, deferred maintenance and off-budget infrastructure financing. Most of these devices exploit the gap between cash flows and economic cost. From a public choice perspective, this gap is an invitation to shift burdens onto future taxpayers who are absent from today's bargaining.

Accrual accounting changes the shadow price of these choices. Take the trade-off between wages and pensions. Under cash, a pay rise costs money immediately, while an actuarially equivalent increase in pension benefits costs nothing in the budget year. Collective bargaining therefore tilts toward deferred compensation. Accrual records the current service cost of the pension in the year it is earned, which makes the two forms of compensation comparable. The accounting change does not dictate the outcome, but it removes a structural bias in favour of hidden deferral.

Asset transactions are reframed even more sharply. Under cash rules, privatisation reduces the reported deficit. Under accrual, selling an asset at fair value simply exchanges one asset for another, so net worth is unchanged unless the price diverges from the asset's value. This shifts the relevant metric of fiscal consolidation from the cash balance to the change in public sector net worth. Comparative work on public sector balance sheets has shown that this broader view can alter assessments of fiscal space and resilience considerably.

The decision relevance of accrual information has limits that advocates sometimes understate. An accrual operating surplus is not a measure of performance, because a government can generate one by under-providing services. Depreciation charges, if poorly integrated into appropriations, may discourage worthwhile capital investment. Evidence also suggests that accrual reporting changes behaviour mainly when it enters the budget itself, as in New Zealand's appropriation system. Statements published in an annual report many months after the fiscal year closes rarely influence the choices they describe.

The implication is that decision value lies in combining measures. Cash remains essential for liquidity management and for gauging the financing requirement relevant to macroeconomic policy. Accrual adds the intertemporal dimension that cash omits. Yet even a complete accrual balance sheet excludes future tax revenues and future social spending, which dominate long-run sustainability. The most informative fiscal frameworks therefore present cash, accrual and long-term projections together. In effect, they approximate intertemporal net worth rather than relying on any single number.

Takeaway

Accounting systems set the shadow price of fiscal choices. Whatever the books fail to record will, sooner or later, become the cheapest way to spend.

Implementation Challenges: Valuation, Pensions, and Institutional Capacity

The first practical obstacle is the opening balance sheet. Many governments lack a complete register of what they own. Their holdings include land acquired centuries ago, military equipment, road networks built in stages and heritage collections with no meaningful market value. Historical cost is frequently unknown, so governments resort to deemed cost, fair value or depreciated replacement cost. These choices are not neutral. Different valuation bases can produce large swings in reported net worth, and subsequent revaluations can overwhelm the operating results that users actually care about.

Pension measurement illustrates how a single assumption can dominate the fiscal picture. The discount rate applied to future benefits determines the reported liability more than almost any other parameter. Many public plans have discounted guaranteed benefits at the expected return on risky plan assets. Financial economics argues instead for a rate that reflects the low risk of the benefits themselves. Novy-Marx and Rauh showed that applying such rates makes US state pension liabilities substantially larger than reported. For obligations of typical duration, a one-point change in the discount rate shifts the liability by well over ten percent.

Institutional capacity is often the binding constraint. Accrual systems require accountants, actuaries and valuers. They need integrated financial management information systems. They also depend on line ministries that understand provisions, impairment and depreciation. Where these are absent, reforms stall into partial adoption, with consultants producing statements that no decision-maker reads. The sequencing literature, associated with practitioners such as Allen Schick, cautions against premature reform. It recommends securing reliable cash control first, then phasing in accrual elements as capacity develops.

Political economy complicates matters further. Accrual reporting exposes liabilities accumulated by previous administrations, so incumbents face asymmetric incentives. They bear the reputational cost of disclosure, while the benefits of better decisions accrue diffusely and slowly. Gaming also migrates rather than disappears. Once cash timing is neutralised, the levers become discount rates, asset useful lives, impairment judgements and the classification of entities inside or outside the reporting boundary. Independent audit offices and fiscal councils therefore become indispensable complements to accrual standards.

Finally, there are the limits of the business model itself. Governments have no ownership interest to measure, raise most revenue without exchange and serve objectives that span generations. Adopting corporate standards uncritically produces figures that look comparable to private sector accounts but mean something quite different. Public sector standards have adapted corporate frameworks, though tensions remain, particularly around social benefits and heritage assets. The aim is accrual information fit for public decisions rather than an imitation of corporate reporting.

Takeaway

Accrual accounting does not end fiscal gimmickry. It relocates gimmickry from the timing of cash to the assumptions behind estimates, which is progress only if someone independent scrutinises those assumptions.

The case for accrual accounting in government rests on a simple proposition: fiscal decisions should face their full economic cost. Cash systems allow costs to be deferred, disguised or shifted to future taxpayers. Accrual systems narrow these opportunities by recognising obligations, consumption of capital and earned revenues when they arise.

Yet the theory also implies modesty. Recognition boundaries encode political judgements. Valuation and discounting assumptions can dominate results. Accrual information changes behaviour only when it is embedded in budgets and appropriations. No balance sheet captures the sovereign power to tax, which is why long-term fiscal projections remain indispensable.

Optimal design therefore treats accrual accounting as part of a broader architecture rather than a destination. That architecture combines cash control for liquidity, accrual measures for intertemporal cost, sustainability projections for unrecognised commitments, and independent oversight of the assumptions that underpin all three. The goal is not to make governments resemble firms. It is to ensure that the real costs of public choices become visible.