It was an easy headline, especially given the report was released on a Sunday. “One in eight Australians live in poverty,” said the ABC. “More than two million Australians in poverty,” ran the AAP copy. Fairfax ran some profiles of people struggling to make ends meet off the back of the Australian Council of Social Services’ new report on the extent of poverty in Australia.

Few bothered to look into the basis for ACOSS’s numbers, beyond that it was an internationally-accepted poverty benchmark. ACOSS used the OECD’s poverty definition, which is 50% of median disposable income. It also provided data for another, less austere benchmark, 60% of median disposable income. On the basis of the lower OECD definition, which ACOSS prefers, 2.265 million Australians are living in poverty. Using the 60% definition, 3.7 million Australians are.

Fifty per cent of median disposable income is $358 a week for a single adult and $752 for a couple with two children.

Neither sum is easy to live on. But they’re the product of an arbitrary benchmark: there will always be a substantial proportion of the population identified as “living in poverty” if you define poverty in relation to median income.

But deprivation-based measures of poverty allow us to move away from arbitrary benchmarks and consider how many people are actually in financial circumstances where they’re unable to afford basics most of us consider essential. And some high-quality work on this has already been done, including by ACOSS.

Earlier this year ACOSS published a report, “Who is missing out? Material deprivation and income support payments“, based on a study by Professor Peter Saunders and Melissa Wong of UNSW. The study, which used data from the 2010 “Poverty and exclusion in modern Australia” (PEMA) survey, looked at how many households displayed “deprivation indicators” based on a range of indicators widely agreed to be “essential” for households.

For example, 99% — or more — of people agreed warm clothes and bedding, medical treatment, ability to afford prescription medicines, and a substantial meal at least once a day were essential. Over 90% agreed up-to-date schoolbooks, the ability for kids to participate in school activities and outings, a decent and secure home, and regular social contact were essential. The only survey indicator that appears to have been widely disputed as “essential” was “a week’s holiday away from home”, which was identified as essential by 54%.

Using median income as a measure of poverty allows us to capture the sense of “relative poverty”, which is important in a prosperous Western country. But the PEMA approach also enables a comparative element. For example, 72% of PEMA respondents agree computer skills are “essential”, even if they would not fit into any meaningful definition of absolute poverty.

Using the PEMA data, Saunders and Wong provided us with an alternative take on poverty in Australia. But the result isn’t one to fill us with complacency: 15.3% of households reported lacking at least 3 of the “essential” indicators.

This included 18.5% who couldn’t afford a week’s holiday; indicators at the other end of the scale are much lower: inability to buy prescribed medicines was 2.9%; inability to afford children’s participation in school activities 2.6%. But 20.7% reported not having $500 in emergency savings; nearly 12% reported being unable to pay a utility bill at least once a year.

The most at-risk group for lacking “essential” items is, unsurprisingly, Newstart-recipient households; 61% reported lacking at least three indicators, although the sample size is too small to provide reliable figures. Households reliant on the disability support pension or parenting payments also featured strongly in households reporting deprivation indicators; age pension recipient households featured much less prominently; indeed, had a lower rate of lacking at least three essential indicators than the community as a whole; it was age pension recipients who didn’t own their own homes who were overrepresented.

This returns us to the debate over whether to increase the level of Newstart assistance. In August, even the Centre for Independent Studies, a serial participant in the “poverty wars”, argued there may be a case for an increase in Newstart for the long-term unemployed, provided it came with stronger requirements for recipients to search for work. This would be cheaper than an across-the-board rise that would simply represent a windfall for the majority of Newstart recipients who only receive it temporarily while between jobs.

And however arbitrary, the weekend report from ACOSS does put a human face on poverty. It’s worth examining to see who else is overrepresented: people in poverty are more likely to be female, more likely to be kids or older people, more likely to be single or lone parents, more likely to be from a non-English speaking country. But the difference between metropolitan and regional/rural poverty is quite low: 12.6% of people in cities and 13.1% of people in regional areas.