Administrative Cost Ratio

Administrative Cost Ratio in Insurers and Health Systems

The administrative cost ratio sets the cost of running a payer, or a whole financing system, against the money that payer handles, showing how much goes on collecting, pooling and paying rather than on care. It is widely quoted in debates about health insurance and health system reform, and easy to misread, because the same label covers several numerators and denominators. The result depends on where administration sits in health system financing and on how the ratio is built, and it is closely linked to the medical loss ratio in United States insurance regulation. OECD comparisons across countries have to be read with those choices in mind, and an illustrative worked example calculates the ratio three ways for one insurer to show why a lower figure is not on its own evidence of better use of resources.

Where administration sits in health financing

Every financing arrangement, whether a national health service, a social insurance fund or a private insurer, has to raise revenue, pool it and purchase care. Each task uses staff and systems that never reach a patient directly, and the ratio measures that overhead relative to the money managed.

The System of Health Accounts (SHA 2011, produced by the OECD, Eurostat and the World Health Organization) draws the boundary used in international health spending data. Its function HC.7 has two classes. Governance and health system administration (HC.7.1) covers planning, policy, regulation and monitoring of the system. Administration of health financing (HC.7.2) covers enrolling members, billing and collecting contributions, risk equalisation, contracting with providers and the claims processing system, including gate-keeping and payments to providers. For private insurers it also covers sales, claim adjudication, actuarial and legal functions, corporate overheads and risk charges.

Two exclusions matter. Administration inside hospitals and practices is counted as part of the care those providers deliver, not under HC.7, and the cost of patients' own paperwork is left out. A system-level ratio built from health accounts therefore measures payer and government administration, not the total administrative effort of the system. The OECD separates three levels, namely financing and governance bodies (macro), provider organisations (meso) and the time clinicians spend on administrative tasks (micro).

Calculating the ratio and the choices that change it

The ratio is a simple quotient, but both parts involve definitional choices that can move the result by several percentage points. Like other financial ratios, it supports comparison only when the parts are defined in the same way. In general form it is written as follows.

$$ACR = \frac{A}{D}$$

where $ACR$ is the administrative cost ratio, $A$ is administrative cost over a period, and $D$ is the denominator for the same period, usually premium revenue for an insurer, total spending by the payer, or current health expenditure for a country.

An insurer with a surplus collects more in premiums than it spends, so the same administrative cost gives a lower ratio on premium than on spending. Removing taxes and regulatory fees from premium first raises the ratio. At country level, payer administration divided by all health expenditure gives a much smaller figure than private insurers' administration divided by their own spending. On the numerator side, the main choices are these.

The OECD has reported that countries value private insurers' administration differently in health accounts. Some report the sum of costs, whereas recommended practice includes profits and brokerage fees. The same insurer can therefore show different ratios in different datasets.

The link with the medical loss ratio

A prominent regulatory use of an administrative share is the medical loss ratio in the United States. The Affordable Care Act requires health insurance issuers to report how premium revenue is spent and to pay rebates when too little goes on clinical services and quality improvement. The rules sit in 45 CFR Part 158, which implements section 2718 of the Public Health Service Act. The ratio is defined as follows.

$$MLR = \frac{C + Q}{P - T}$$

where $MLR$ is the medical loss ratio, $C$ is incurred claims for clinical services, $Q$ is expenditure on activities that improve health care quality, $P$ is earned premium, and $T$ is federal and state taxes and licensing and regulatory fees. The regulation also adjusts the denominator for risk adjustment, risk corridors and reinsurance, applies a credibility adjustment where one is due, and rounds the result to three decimal places.

Under 45 CFR 158.210, an issuer must pay a rebate if its ratio is below 85% in the large group market or below 80% in the small group and individual markets. A state may set a higher standard under 45 CFR 158.211, and the individual market standard can be adjusted for a state. Under 45 CFR 158.240 the rebate equals premium after taxes and fees multiplied by the gap between the required and achieved ratio. Medicare Advantage organisations face a separate 85% requirement in 42 CFR 422.2410.

The complement of the medical loss ratio is often called the administrative share, but it contains more than administration. Rearranging the definitions gives the following expression.

$$1 - MLR = \frac{N + S}{P - T}$$

where $N$ is non-claims costs other than taxes, fees and quality improvement, $S$ is the insurer's underwriting gain or margin, and the other terms are as defined above. The complement mixes administration with profit, so a saving on administration that is kept as margin leaves the medical loss ratio unchanged.

The regulation also decides which side of the line some activities fall on. Under 45 CFR 158.150, activities designed primarily to contain costs, all retrospective and concurrent utilisation review, fraud prevention, provider network contracting and marketing cannot count as quality improvement; they are reported as non-claims costs, which 45 CFR 158.160 lists as including cost-containment expenses, broker fees and general administration. Prospective drug utilisation review aimed at adverse drug interactions is an exception that counts as quality improvement, and claims recovered through fraud reduction, up to the amount spent on it, stay in incurred claims.

International comparisons of administrative spending

At the level of whole systems, the main comparable source is OECD health expenditure data built on the health accounts framework. In its 2017 report on wasteful spending, the OECD found that member countries spent an average of around 3% of health spending on administration in 2014. The share was double that level in France and higher still in the United States, while several countries reported less than half the average.

The OECD linked these differences to how financing is organised. Single-payer systems, whether the payer is a government or a single social health insurance fund, tended to spend less on administration than multi-payer systems, and systems where people choose their insurer tended to spend more than those with automatic affiliation. Among countries with competing insurers, the figure ranged from 3.3% of public health spending in the Czech Republic to more than 5% in Germany. In 2014 or the nearest year, voluntary private insurance schemes spent from about 9% of their spending on administration in Australia to more than 30% in Spain, the United Kingdom and Austria, whereas public schemes spent at most 9% in any country.

The OECD also warned that countries record administrative spending in different ways, and that simple international comparisons reflect differences in governance and financing structures and show only the costs of administration, not its benefits.

Why a low ratio does not prove efficiency

A low ratio is often read as evidence of an efficient payer, but it sets one input against a scale variable and says nothing about what the administration achieves. Technical efficiency concerns outputs obtained from given inputs, and a payer's outputs include accurate payment, control of inappropriate use, protection against fraud and the information needed to purchase care well. Administration that cuts claims by more than it costs lowers total spending while raising the ratio.

The OECD makes the same point, noting that administration is not in itself waste and that its costs should be weighed against benefits that are hard to measure. Its examples include pay-for-performance schemes, which need extra reporting, and health technology assessment, which costs money but supports better informed coverage decisions. The same reasoning applies to the running costs of utilisation review, fraud control and value-based purchasing, which are usually counted as administration but intended to change the size or mix of claims. Under the United States rules, quality-linked incentive payments to providers themselves count as incurred claims; only the cost of running such schemes is administration.

The ratio also moves with scale and price. Two insurers with the same administrative cost per member will show different ratios if one covers a sicker or more expensive population, and a rise in provider prices lowers the ratio with no change in administration. Administrative cost per member is a useful companion measure for this reason.

Worked example with one insurer, three ratios and a cut in claims control

The figures below are illustrative and describe a hypothetical insurer in the United States individual market over one year. Earned premium is $500 million, taxes and regulatory fees are $20 million, incurred claims are $382 million, quality improvement is $10 million, and other administration (claims processing, utilisation review, fraud prevention, broker commissions and overheads) is $60 million. The margin is $28 million, since 500 − 20 − 382 − 10 − 60 = 28. For simplicity the example uses a single year, although 45 CFR 158.220 aggregates three years of data, and it assumes fully credible experience and ignores risk programme adjustments and the fraud recovery adjustment.

1. Ratio on premium revenue. Administrative cost is divided by premium.

$$ACR_P = \frac{60}{500} = 0.12$$

where $ACR_P$ is the ratio on premium revenue, so 12.0% of premium goes on administration.

2. Ratio on total spending. Claims, quality improvement and administration total $452 million (382 + 10 + 60 = 452), which gives the following ratio.

$$ACR_S = \frac{60}{452} \approx 0.133$$

where $ACR_S$ is the ratio on total spending, about 13.3%.

3. Medical loss ratio and its complement. The regulatory definition gives the following.

$$MLR = \frac{382 + 10}{500 - 20} = \frac{392}{480} \approx 0.817$$

where the terms are as defined in the medical loss ratio section. At 81.7% the insurer meets the 80% individual market standard, so no rebate is due. The complement, 18.3%, is administration plus margin over adjusted premium, since (60 + 28) / 480 ≈ 0.183. Three reasonable calculations give 12.0%, 13.3% and 18.3% for the same insurer.

4. The same insurer in the large group market. Against the 85% standard the shortfall is 0.850 − 0.817 = 0.033, so the rebate would be as follows.

$$R = 480 \times 0.033 = 15.84$$

where $R$ is the rebate in millions of dollars, $15.84 million.

5. A cut in claims control. Returning to the individual market, suppose the insurer cuts utilisation review and fraud prevention by $5 million and, as a result, pays $12 million more in claims that would otherwise have been stopped. Administration falls to $55 million and claims rise to $394 million, and the ratios are recalculated as follows.

$$ACR_P = \frac{55}{500} = 0.11, \qquad MLR = \frac{394 + 10}{480} = \frac{404}{480} \approx 0.842$$

where the terms are as defined above. The administrative ratio falls from 12.0% to 11.0% and the medical loss ratio rises from 81.7% to 84.2%, both apparent improvements. Yet spending on claims, quality improvement and administration rises from $452 million to $459 million and the margin falls from $28 million to $21 million, so covering the same members now costs $7 million more.

6. The same cut under the large group standard. At 84.2% the insurer would still fall short of 85%, so the rebate is recalculated as follows.

$$R = 480 \times (0.850 - 0.842) = 480 \times 0.008 = 3.84$$

where $R$ is again the rebate in millions of dollars. The margin after rebate rises from $12.16 million (28 − 15.84 = 12.16) to $17.16 million (21 − 3.84 = 17.16), even though total cost has risen by $7 million. Under a medical loss ratio floor, cutting claims control can therefore improve the insurer's net position while making the cover more expensive.

How the ratio relates to nearby measures

Several measures overlap with the administrative cost ratio and are sometimes used interchangeably. They differ in what they divide and what they are for.

Unlike measures of allocative efficiency or cost-effectiveness, the ratio says nothing about whether the care purchased is good value.

Common misreadings of the ratio

A common error is to compare ratios built on different definitions, such as a ratio on gross premium against one on total spending, or a figure that includes profit against one that does not; one minus the medical loss ratio, for instance, also contains the insurer's margin. A low ratio is read as efficiency when it may reflect high claims, high prices or too little claims control, and international rankings are read as measures of total administrative effort although they exclude administration inside hospitals and practices. Finally, the ratio is sometimes treated as a quality indicator. It shows where money goes, not the quality of care or the health it produces, so it informs judgements about health system performance only alongside measures of access, quality and financial protection.

Sources

• Centers for Medicare and Medicaid Services. Medical Loss Ratio. CMS Center for Consumer Information and Insurance Oversight; accessed 2026. https://www.cms.gov/marketplace/private-health-insurance/medical-loss-ratio
• Code of Federal Regulations. Title 42, section 422.2410: Medicare Advantage medical loss ratio, general requirements. Current as at 2026. https://www.law.cornell.edu/cfr/text/42/422.2410
• Code of Federal Regulations. Title 45, Part 158: Issuer use of premium revenue: reporting and rebate requirements, sections 158.101, 158.130, 158.140, 158.150, 158.160, 158.210, 158.211, 158.220, 158.221 and 158.240. Current as at 2026. https://www.ecfr.gov/current/title-45/subtitle-A/subchapter-B/part-158
• OECD. Tackling Wasteful Spending on Health. Paris: OECD Publishing; 2017. Chapters 1 and 6. https://doi.org/10.1787/9789264266414-en
• OECD, Eurostat, World Health Organization. A System of Health Accounts 2011: Revised Edition. Paris: OECD Publishing; 2017. Chapters 4 and 5. https://doi.org/10.1787/9789264270985-en