Opinion & Commentary
Poor way to detect poverty
Someone at the Australian Council of Social Service (ACOSS) is forgetting their Aesop. You will recall the little boy who kept pretending there was a wolf attacking the sheep. When eventually a real wolf appeared, nobody believed him, for they had grown tired of his false alarms and exaggeration.
ACOSS regularly releases alarmist reports about poverty in Australia. Its latest claims that two million Australians – 10 per cent of the population – are living in poverty. It adds that if we adopted the poverty measure favoured by the Brits and Irish, this figure would double to one in five.
Measuring the size of the poverty problem is much like determining the length of a piece of string, for poverty estimates vary wildly according to the definition used. ACOSS defines ‘poverty’ as household income below 50% of the median income. But it could equally have drawn its line at 60% (the preferred measure in the EU), 40%, or any other proportion. There is no rationale to these things – it really is totally arbitrary.
A moment’s reflection also makes clear that what is being discussed here is not ‘poverty,’ but income inequality. Comparing the incomes of people at the bottom of the distribution with those higher up tells us about the difference between them, but it tells us nothing about whether they are ‘poor’ or ‘rich.’ There is a broad consensus among those who study these things that ‘poverty’ has to do with people’s inability to maintain an adequate living standard. But income is a very poor guide to people’s standard of living.
People’s incomes fluctuate over time, as they change jobs, retire, start a family, begin a college course, get a promotion, become redundant, go part-time, or whatever. I might be under the ACOSS poverty line today, but above it tomorrow (indeed, when I was younger, I was under this line). Research following a panel of Australian households over several years found that 12% of them had less than half the median income in one year, but only 6% had an income this low for two years running, and just 4% stayed under the line for three years.
Households tend to save when their income is high, borrow when it is low, and adjust their spending to flatten out the peaks and troughs as they go through life. Their living standards therefore remain much more stable than their incomes. Young people often spend more than they earn – few of us would ever buy a house if we didn’t. Older people often live off assets bought earlier in life, when their earnings were higher. For all these reasons, taking a simple snapshot of people’s incomes at just one point in time, as ACOSS does, is a very bad way of assessing their living standards.
Recognising this, researchers have started using other possible measures of poverty. A few years ago, the Department of Family & Community Services in Canberra came up with measures of what it called ‘financial stress.’ It counted how many people could not afford to pay their heating bills, buy a holiday away from home, have a regular night out. On indicators like these, it found just 3% of households were suffering ‘multiple hardships.’ It also found that only a small proportion of households with low incomes were deprived – low income was a bad predictor of hardship. It concluded: ‘A simple focus on levels of income would appear to be misplaced.’
Researchers at the Melbourne Institute have come to the same conclusion. The Institute’s Gary Marks reports that low incomes correlate very poorly with other measures of ‘poverty’ such as financial stress and ‘subjective poverty’ (the belief that you are poor). And his colleague, Bruce Headey, has compared households’ incomes with their consumption patterns and found considerable divergence. In turns out that having a low income need not mean you buy less food, clothing, transportation, gas and electricity, health insurance, alcohol, meals out or home maintenance than other people do.
Headey suggests combining income and consumption into a single measure of poverty. Measuring income alone, he finds 12% of Australians are ‘poor’ (an income less than half the median), but when their consumption patterns are included, this falls to just 3% (where income and consumption are both less than half the median). When he then analyses this new measure over time, he finds fewer than 1% remain ‘poor’ over two successive years.
Headey concludes: ‘Existing income-based measures are seriously in error. The results they give are much too high.’
ACOSS should take heed of research findings like these. Wild claims that one in ten, or even one in five, Australians is ‘poor’ are guaranteed to get them headlines, but the estimates on which these claims are based are meaningless and absurdly exaggerated. One day the Australian public will rumble this, and when they do, the poverty lobby in this country will lose all remaining credibility.
Professor Peter Saunders is the Social Research Director at The Centre for Independent Studies.
