Why Can the Same CPI Number Be Used to Index Benefits but Not Represent Your Cost of Living?

Geoffrey Chen

When Numbers Start Making Decisions · Season One, “People on Either Side of a Threshold” · Article 2

1. The same adjustment enters different lives

Imagine two people receiving JobSeeker Payment. They are the same age and have the same family status, so the same maximum basic rate applies to them.

One rents a home alone. Most of her income goes towards rent, food, public transport and prescription medicines. The other lives with family and does not pay market rent directly, but has to meet the costs of a car, insurance and a long commute. When social security payments are indexed to the Consumer Price Index, both receive the same percentage increase in their basic payment.

The institutional logic is easy to understand. A payment category needs a uniform rule. The government cannot construct a separate price index for every recipient.

But when these two people open their bank accounts, the effect may feel entirely different. If rents and medicines have risen faster during the adjustment period, the first person’s essential expenses may be squeezing her even harder. If fuel, repairs and insurance have risen more quickly, the second person may face the greater pressure.

A familiar argument follows. One side says, “The CPI already measures inflation, and the payment has been adjusted as the law requires.” The other says, “That number bears no resemblance to my cost of living.”

Both statements may be true.

The CPI can perform an aggregate measurement accurately, and a social security system can reasonably use that measurement to create a common adjustment rule. That does not mean the CPI describes every household. Nor does a CPI-linked increase prove that recipients can maintain their previous standard of living.

The important question is not whether the CPI is “real”. It is what new decision-making function the number acquires when it moves from a statistical table into a legal formula.

2. The CPI does not measure a “typical household”

The Australian Bureau of Statistics defines the CPI as a general measure of price change for goods and services purchased by Australian households. Food, housing, transport, health, education and other expenditure are organised into eleven broad groups and 87 expenditure classes. Each class receives a weight based on its relative share of household-sector expenditure, and the national result is a weighted average of the indexes for the eight capital cities. ABS: Consumer Price Index, Australia methodology

This method is often simplified as a “representative basket”. That image is useful, but it can encourage us to imagine a household that does not exist: an average family that pays rent, buys a new home, visits doctors, drives, studies and travels in exactly the national proportions, while the statistical agency watches how much more its monthly shopping costs.

In reality, CPI weights reflect the expenditure structure of all in-scope households together. They are not the complete budget of an average family. The ABS points out that the CPI basket includes both rent payments and the cost of new dwellings bought by owner-occupiers, although an individual household would not normally incur both forms of housing expenditure at once. The price experience of any particular household is therefore unlikely to match the national CPI exactly. ABS: Frequently asked questions about the Consumer Price Index

This is not a defect in the statistical design. It is a basic property of an aggregate indicator. To answer how prices are changing for the Australian household sector as a whole, the statistical system must compress millions of different consumption patterns into one number that remains comparable over time.

That compression involves at least three institutional choices: which households to measure, which goods and services to include, and how much weight to give each kind of expenditure. In practice, the Australian CPI’s reference population is private households in the eight capital cities. It provides an overall account of price change in the main population centres, but it may not capture the prices faced in rural and remote areas. Nor does it tell us whether the absolute price level is higher in Sydney than in Hobart. Each city index measures how its own prices change over time, not the difference in price levels between cities.

The CPI therefore answers this question:

According to the expenditure weights of the target household population as a whole, by how much has the price of a basket of goods and services changed?

It does not directly answer:

How much more must this particular household spend to maintain its previous way of life?

The questions are related, but they are not the same.

3. Why can an aggregate number still be used to index social security payments?

A social security system cannot merely promise to “increase payments appropriately when prices rise”. It has to specify the index, the comparison periods, the effective date, the rounding method and what happens if prices fall. Without those rules, every adjustment could become a new ad hoc negotiation.

In Australia, common benefit rates including JobSeeker Payment, Special Benefit and partnered Parenting Payment are generally indexed to CPI movements. Adjustments take effect on 20 March and 20 September each year. The March indexation uses the CPI change from the preceding June to December, while the September indexation uses the change from the preceding December to June. The governing framework comes from Part 3.16 of the Social Security Act and its associated calculation rules. Department of Social Services: Common benefit rates

The CPI is suitable for this role not because it knows what each recipient buys, but because it has properties that an administrable rule needs:

  • an independent statistical agency produces it continuously under a published method;
  • it covers a broad range of goods and services regularly purchased by households;
  • it provides a consistent national adjustment rate without examining every recipient;
  • it can be calculated repeatedly across periods, allowing governments, recipients and reviewers to understand how the formula operates; and
  • it reduces the scope for the real value of payments to be frozen arbitrarily or determined only by short-term political judgement.

Social security law is not treating the CPI as the personal budget of every recipient. It is using the index as a common measuring stick. The law does not require a perfect match with each household; it requires a benchmark broad, stable, transparent and executable enough to support a uniform rule.

An important transformation occurs here. In the hands of the ABS, the CPI is descriptive: it reports how prices have moved. Inside a social security formula, it begins to have normative consequences: the law uses that reported movement to determine how a payment should change.

The number did not make that decision by itself. Legislation and policy chose to place it in the formula.

4. Why can the same percentage not represent everyone?

The most obvious reason is that different households spend their money in different places.

Older households may devote a larger share of expenditure to health. Families with children face education and care costs. Renters experience changes in rent, whereas outright homeowners pay neither rent nor mortgage interest. Remote households may depend more heavily on cars while also facing fewer retailers and higher transport costs. Even two households with the same income can have sharply different price experiences because their housing, health, family structure and location differ.

The ABS’s Selected Living Cost Indexes make these differences visible. In the weighting pattern published for March 2026, health represented about 12.41 per cent of expenditure for age pensioner households, compared with about 6.73 per cent in the CPI. Housing represented about 25.02 per cent for other government transfer recipient households, compared with 21.39 per cent in the CPI. Insurance and financial services represented about 18.69 per cent for employee households, compared with 5.58 per cent in the CPI. These figures do not imply that every pensioner, transfer recipient or employee spends in those proportions. They show that expenditure patterns differ substantially even when households are divided into broad groups. ABS: Selected Living Cost Indexes, Australia methodology

The measurement concepts also differ. The CPI mainly uses an acquisitions approach: it measures the prices of goods and services acquired by the household sector and serves as a macroeconomic indicator of inflation. A living cost index is closer to an outlays approach, focusing on the expenses particular household groups actually pay to obtain goods and services.

Housing finance provides the clearest example. The CPI includes the cost of new dwellings purchased by owner-occupiers but excludes mortgage interest. The living cost indexes exclude the purchase of the dwelling itself but include mortgage interest and consumer credit charges.

When interest rates change rapidly, households with mortgages may experience a severe increase in cash-flow pressure, even though that pressure does not enter the CPI in the form of mortgage interest. The CPI has not accidentally forgotten an obvious bill. It measures a different concept. The apparent contradiction arises when a macroeconomic price-inflation measure is treated as if it were a household cash-outlay measure.

Cost of living also involves the ways households respond to changing prices. A household may switch to cheaper food, postpone dental treatment, stop heating the home, or move further away from work. If its recorded expenditure does not rise in the same proportion as prices, that does not prove its standard of living remained unchanged.

The ABS also explains that a strict cost-of-living measure would need to consider the minimum expenditure required to maintain a given standard of living. That standard depends not only on prices, but also on health, surroundings, family responsibilities and social participation. No single official index can completely measure the welfare of every household.

Saying “my cost of living rose faster than the CPI” is therefore not necessarily an allegation that the statistics are wrong. It may simply mean that the household’s expenditure weights, location, debt structure and essential consumption differ from the aggregate basket.

5. Between “price change” and “payment adequacy” lies an additional judgement

Debates about benefit indexation often join three distinct propositions:

  1. How much did household prices generally rise?
  2. By what percentage should a social security payment be adjusted?
  3. Is the adjusted payment sufficient to support an acceptable life?

The CPI measures the first proposition directly. Social security law borrows the CPI and makes an institutional decision about the second. The third cannot be derived automatically from the first two.

If a payment was already below the level required to meet basic needs before indexation, increasing it exactly in line with the CPI will approximately preserve its previous relationship to general prices. Indexation can help prevent inflation from continuously eroding an established amount, but it does not prove that the original amount was adequate.

Even when the CPI is perfectly accurate, adjustment can lag behind current experience because of timing. Changes on 20 March and 20 September use CPI movements observed over earlier periods. Prices may already have changed for recipients, while the payment waits for the statutory adjustment date. A uniform scheme must choose observation periods and effective dates, but the administrative clarity of those choices also creates a delay.

“The payment has been indexed to the CPI” proves only that the formula operated. It cannot, by itself, prove that a policy objective was achieved. If the objective is simply to make payments follow general price movements, CPI indexation is strong evidence. If the objective is to ensure that a particular household group can afford a defined set of essentials, evidence about that group’s actual expenditure and the adequacy of the payment is also needed.

This is where an indicator most easily exceeds its authority after it gains decision-making power. A measure designed to answer “how much did prices change?” is quietly asked to answer “is this amount now enough?” The first is a question of price measurement. The second includes social judgements about minimum living standards, dignity and public responsibility.

No price index can make that value choice for society.

6. If the CPI does not match my life, can I appeal?

An individual generally cannot demand a personal recalculation of a nationally indexed benefit rate simply because their rent rose faster than the CPI. The CPI has not made a factual finding about that individual and does not claim to reconstruct their private budget. Where the law applies the same index to a category of basic rates, a difference between an individual’s spending pattern and the index is not ordinarily a calculation error.

This differs from a credit score, a benefit eligibility assessment or a tax decision. Those systems use personal information to decide something about a specific person, so the person must be able to inspect the information, correct errors and seek review. CPI indexation adjusts a class of amounts collectively. Its correction mechanisms are correspondingly different: transparent statistical methods, regularly updated weights, clear descriptions of geographic and conceptual coverage, preserved historical data, methodological scrutiny, and a legal rule that identifies the relevant index period.

The absence of a personal appeal against the national CPI does not mean the system needs no review. Review should occur where the number is connected to its purpose:

  • How far does the expenditure pattern of the payment’s recipient group differ from that of the overall CPI household population?
  • Have general inflation, actual living expenses and payment adequacy been mistakenly treated as the same issue?
  • What lag is created by observation periods, rounding and commencement dates?
  • Are supplementary policy tools available when essential expenses such as housing, energy or health undergo exceptional shocks?
  • Is the underlying basic rate periodically reviewed outside the indexation formula?

Australia’s pension system already recognises that one index may not be enough. The maximum basic rates of several major pensions are indexed by the greater movement in the CPI or the Pensioner and Beneficiary Living Cost Index, and are then protected by a benchmark linked to Male Total Average Weekly Earnings. The PBLCI more closely reflects the expenditure structure of pensioner and other government transfer recipient households, while the earnings benchmark reflects the relationship between pensions and general community income. Department of Social Services: Common provisions affecting indexation of pensions

This arrangement does not prove that pensions have reached an uncontested level of adequacy. It reveals an important principle: when a number is used to protect people’s living standards, a system can consider general prices, the living expenses of particular groups and changes in society-wide earnings at the same time. It need not delegate the whole judgement to a single indicator.

Conclusion: the CPI can be a common measuring stick, but it cannot judge lived experience

The CPI never promised to describe your household. It combines broad price changes faced by households in the eight capital cities into a national indicator, providing a common language for macroeconomic policy, contracts and indexation. Its transparency, continuity, breadth and repeatability make it reasonable for a social security system to use the CPI when adjusting certain payments.

But that reasonableness has clear limits.

Indexing a payment to the CPI does not mean every essential expense faced by a recipient rose at the same rate. It does not erase differences in geography, housing and family structure. Nor does it establish that the adjusted amount can sustain a socially accepted standard of living.

The sound conclusion is therefore neither “the CPI does not match me, so it is false” nor “the payment rose with the CPI, so the cost-of-living problem has been solved”. A more accurate judgement is:

The CPI can coordinate a national price adjustment, but it cannot independently determine whether every person’s life is adequately protected.

Modern institutions must compress different lives into common rules. Without doing so, they could not make stable and predictable public decisions. Once a number moves from description to decision, however, the institution must explain why it suits the purpose, which experiences are excluded, and what supplementary judgement remains necessary.

The power of a number should not exceed its capacity to measure. The consequences of a formula should not exceed the institution’s capacity to explain, review and correct that formula.

Primary sources

Discover more from Geoffrey Chen

Subscribe to get the latest posts sent to your email.