# Per-capita and per-unit costs: comparing apples across agencies
A county in Ohio spends $340 million a year running its government. A mid-sized city in Germany spends €280 million. Which one is more efficient? You cannot answer that question with the raw totals. You need to divide by something that makes the numbers comparable: population, caseload, or service units. That normalization is the entire craft of this lesson.
Total spending is driven mostly by size. A county of 2 million people will always spend more than a county of 200,000, regardless of how well either is run. Comparing totals across jurisdictions (the general term for any governing body: a county, city, state, or national agency) tells you almost nothing about efficiency or service quality.
The fix is to normalize: divide total cost by a denominator that reflects demand or scale.
Three denominators matter most in public sector finance:
$$\text{Per-unit cost} = \frac{\text{Total program or agency expenditure}}{\text{Denominator (population, cases, or units)}}$$
A US county spends $85 million annually on its sheriff's department (the county-level law enforcement agency) and has a population of 410,000.
$85,000,000 / 410,000 = $207 per resident per year
Now compare a European city of 620,000 residents spending €95 million on municipal policing:
€95,000,000 / 620,000 = €153 per resident per year
Already more comparable than raw totals, but currency and cost-of-living differences still distort the picture (more on that below).
Population-wide averages hide huge variation when a program serves a small subset of residents. If you're evaluating child welfare services, unemployment benefits, or public housing, use the number of *cases* or *clients*, not total population.
Worked example: child protective services (CPS)
A county spends $22 million on CPS and handles 4,400 active cases in a year.
$22,000,000 / 4,400 = $5,000 per case
This figure is what you'd use to benchmark against a neighboring county, or against a national average like those tracked by the US Administration for Children and Families (part of the US Department of Health and Human Services).
For operational services, use physical output units. This is the closest public finance gets to a unit economics view familiar from corporate finance.
Common per-unit metrics:
| Service | Denominator | Typical use |
|---|---|---|
| Waste collection | Cost per ton collected | Benchmarking contractor bids |
| Public transit | Cost per passenger-mile or per revenue-hour | Transit agency efficiency |
| KKThe average number of new users each existing user generates through referrals. Above 1.0, growth compounds on itself and becomes exponential.Voir la définition complète →-12 education | Cost per pupil (enrolled student) | State and district comparisons |
| Road maintenance | Cost per lane-mile | Infrastructure benchmarking |
| Hospitals (public) | Cost per bed-day or per discharge | Health system efficiency |
Worked example: cost per pupil
A US school district spends $410 million and enrolls 38,000 students.
$410,000,000 / 38,000 = $10,789 per pupil
As of recent estimates, US average per-pupil spending is roughly $15,000 to $16,000 (figures vary by state; see the National Center for Education Statistics for official state-by-state data). A district at $10,789 is running notably lean, or under-resourced, depending on outcomes.
Comparing a US county in dollars to a European municipality in euros requires more than a currency conversion. A euro buys different amounts of local goods and labor in Lisbon versus Munich, just as a dollar buys different amounts in rural Mississippi versus San Francisco.
Purchasing Power Parity (PPP) is an exchange rate adjustment that reflects relative local buying power rather than market currency rates. The OECD and World Bank publish PPP conversion factors annually (see OECD PPP data).
Why this matters for benchmarking:
A German city spending €153 per resident on policing and a US county spending $207 are not directly comparable by market exchange rate alone. If the market rate is roughly $1.08 per euro (a rate that fluctuates and should be checked at time of use), the €153 converts to about $165, market-rate. But if PPP adjustment shows the euro has more local purchasing power in that region (that is, labor and services cost less locally than the exchange rate implies), the real comparable figure could be higher or lower once adjusted. Always state which method you used: market exchange rate or PPP.
A common distortion in public reporting: agencies choose denominators that flatter their numbers.
Always ask: what exactly is in the denominator, and is it the same definition used by whatever benchmark you're comparing against?
Vérification des acquis
1. Why is comparing raw total budgets between a large county and a small county misleading?
2. A city wants to evaluate the efficiency of its public library system. Which denominator would best isolate whether the system is efficient relative to actual usage, rather than just population size?
3. Two agencies serve populations of very different sizes. Agency A has a lower per-capita cost than Agency B. What can you correctly conclude?
4. Select ALL correct answers about choosing the right denominator for normalizing costs.
Sélectionnez toutes les réponses correctes.
5. Select ALL correct answers about the limitations of per-capita comparisons across countries.
Sélectionnez toutes les réponses correctes.
Before citing a per-capita or per-unit figure in a board memo or analysis, check:
1. Same fiscal year? Budgets shift year to year; mismatched years distort trend comparisons.
2. Same currency basis? Market rate or PPP, and is that stated?
3. Same denominator definition? Population served versus total population, enrolled versus attending, gross versus net cases.
4. Same cost scope? Does the total include capital expenditurecapital expenditureCapital Expenditure (CapEx) is money spent to acquire, upgrade, or extend long-lived assets like equipment, property, or software that deliver value over multiple years.Voir la définition complète → (one-time infrastructure spend) or only operating expenditure (recurring costs)? Mixing the two skews comparisons badly.
5. Inflation-adjusted? Comparing 2019 dollars to 2026 dollars without adjustment overstates real spending growth. Use a standard deflator like the US Bureau of Economic Analysis price indices or Eurostat's HICP (Harmonised Index of Consumer Prices) for Europe.
Worked example: real growth check
A county's per-capita spending rose from $180 in 2020 to $207 in 2025. Nominal growth: 15%. If cumulative inflation over that period was roughly 22% (a plausible US estimate for that window; always check current CPI data), real per-capita spending actually *fell*, meaning the county's purchasing power per resident declined even as the sticker figure rose.
🎬 [VIDEO: "Purchasing Power Parity Explained" - https://www.youtube.com/results?search_query=purchasing+power+parity+explained - a short primer on how PPP adjustments work and why they matter for cross-country comparisons]