Government Output Measurement

The government output measurement problem arises because public-sector services—defence, education, healthcare, administration—are not sold on markets and have no observable prices. National accountants resolve this by valuing government output at its cost of production (wages, materials, capital depreciation), not at what users would willingly pay. This accounting choice has three major consequences: it inflates measured GDP, it obscures whether public services are becoming more or less efficient, and it creates distortions in international comparisons.

The fundamental measurement problem

When the market sells a haircut for $30, the hairdresser’s output is worth $30. When the supermarket spends $8 on cashier wages and $2 on materials to stock a shelf, and a customer buys items for $50, the supermarket’s output value is $50 (the retail markup). In both cases, the market price reveals what consumers are willing to pay, and national accounts can use that price as a proxy for real output.

Government provides no such signal. A public school does not charge tuition; a court hearing has no admission fee; national defence cannot be priced per unit of protection. Yet these are real goods and services that people consume and value. How should national accountants measure their output?

The answer adopted by most countries, including the US, the UK, and those following the System of National Accounts (SNA), is to value government output at its cost of production. If a school district spends $500 million annually on teachers, buildings, materials, and administration, that $500 million is recorded as the “output” of the education sector in national accounts. If a hospital spends $2 billion on nurses, doctors, equipment, and buildings, that $2 billion is its output. By definition, government output grows whenever government spending grows—a logical tautology that has real consequences.

Why input cost creates distortions

Invisible efficiency: Suppose a school district trains teachers to teach more students per classroom through better pedagogy, halving the cost per pupil. Output, measured by input cost, has just fallen—the opposite of what actually happened. The school system is more efficient, but national accounts register it as a contraction. Conversely, if inefficiency doubles spending whilst teaching fewer students, the measured output rises. This perverse accounting invisibly rewards waste and penalises productivity.

GDP inflation without welfare gains: When government does more—hires more teachers, builds more hospitals, or raises military spending—nominal GDP automatically rises by the full amount spent, regardless of whether the services actually improve. A country could hire millions of people to dig holes and fill them in, and if government pays for it, GDP grows. This matters because policymakers and the public often use GDP growth as a proxy for wellbeing, whilst ignoring that quality improvements in education or healthcare would be invisible (or even counted as a contraction) under input-cost measurement.

International comparison distortions: Countries differ in how much they rely on public provision versus private markets. The Nordic countries, with large public sectors providing healthcare, education, and childcare, record these as government output at input cost. The United States, where private insurance and private schools are more prevalent, records the same services as private-sector output with market prices. All else equal, a country shifting from private to public provision would show lower measured productivity even if the actual services remained identical. This makes cross-country GDP comparisons less meaningful than they appear.

Alternative approaches and their trade-offs

The input-cost method (current standard) is easy to implement: just tally government outlays. But it embeds the assumption that more spending always means better output, which is false.

Output indicators attempt to measure quality directly: student test scores, hospital readmission rates, crime solved per police officer, and so on. If a school district spends the same but improves test scores, this method would record output growth. But indicators vary by sector, are often contested, and create incentives to game the metrics (teaching to tests, rejecting sicker patients). Furthermore, aggregating diverse indicators (test scores, crime rates, life expectancy) into a single “government output” index requires arbitrary weighting.

Market-equivalent pricing asks: what would this service cost if sold privately? If government provides healthcare at cost £4 billion per year, a private insurance market might charge £5 billion for equivalent care. Using the market price inflates the measured government output. But “equivalent care” is slippery, and this method ignores that government services may differ in scope or quality from private alternatives.

Capodanno or volume-based methods measure output by physical units (teacher-student hours, hospital bed-days, court cases resolved) rather than input cost or price. This is intuitive—more students taught means more output—but assumes constant quality and ignores shifts in service mix or complexity.

The System of National Accounts (SNA) has gradually moved toward output indicators where feasible, recognizing input-cost’s flaws. A 2008 revision of the SNA encouraged countries to adjust government output estimates by indexing them to concrete quality measures when data permit. But widespread adoption remains patchy: most countries still lean heavily on input cost for lack of better alternatives.

The real-world stakes

Government output typically comprises 15–25% of GDP in developed economies. If this component is measured with bias, the headline growth rate is skewed. Suppose “true” government sector productivity rises 1% per year (measured by output indicators) whilst input cost grows 2%. Under input-cost accounting, government output grows 2%, pulling overall GDP growth 0.3 percentage points higher than the true figure (if government is 15% of GDP). Over decades, this compounds into a systematic overestimation of national progress.

Countries with larger public sectors are hit harder. Scandinavia, with public sectors near 30% of GDP, sees a larger fraction of its “growth” driven by input-cost accounting noise than countries with smaller public sectors. This does not imply Scandinavian growth is false—public services may genuinely be high-quality—but it does mean comparisons with market-driven economies are hazardous.

In practical policy terms, input-cost measurement can obscure the case for public-sector reform. If measured output rises whenever spending rises, policymakers lack a clear signal when services are deteriorating. A shrinking school system measured as declining output (falling input cost) might actually be more efficient. Conversely, a swelling bureaucracy shows as growth.

The OECD and emerging consensus

The OECD has pushed members toward more sophisticated government output methods, publishing guidelines for quality adjustment. Some countries have made progress: the UK’s Office for National Statistics (ONS) now adjusts education output by test-score improvements and healthcare output by survival rates. But these are exceptions. Most nations still default to input-cost estimates, partly because they are simple and partly because strong output indicators are not always available (e.g., what measures the “output” of a central bank or foreign ministry?).

A growing intellectual consensus agrees that input-cost measurement is inadequate but that no perfect alternative exists. The practical path forward is sector-by-sector adjustment: use output indicators where data exist (education, healthcare, transport), accept input-cost estimates for complex functions (defence, justice, diplomacy), and acknowledge the uncertainty in published figures. National accounts now increasingly publish “memo items” showing both input-cost and alternative estimates, inviting users to choose.

See also