The Smith Family's poverty report is simply misleading

In January we published a paper, Poor Arguments: A Response to the Smith Family Report on Poverty in Australia, in which we claimed that a report on poverty produced last November by the Smith Family was misleading. In particular, we questioned the two central claims of that report – that one in eight Australians is living in poverty today, and that poverty has worsened over the last ten years. We believe that neither of these claims can stand up to critical scrutiny, although both attracted widespread media attention at the time the report was published.

Our paper attracted a lot of media interest (not least in Kelly Burke’s article, ‘Poverty research claims under fire’, in The Age). Radio, television and newspapers were all keenly interested in what some saw as our ‘attack’ on a worthy charity, and the Smith Family responded to our critique with a vigorous defence of its original claims. So where do things stand now that the dust has started to settle?

There is clearly some common ground between us. Everybody accepts, for example, that ‘absolute poverty’ – an income insufficient to buy nutritious food, clothe yourself properly, and get a roof over your head – is very rare in Australia. Everybody also accepts that even the poorest people in Australia are better off today than they were ten years ago – the Smith Family’s own figures show that the poorest individuals increased their incomes by an average of $38 per week over this period in real terms (i.e. after allowing for inflation).

But wait a moment! If the ‘poor’ got better off over the last ten years, how could the Smith Family conclude in its report that poverty had worsened? The answer is that it was talking about ‘relative poverty’, and that the way it measured this depended less on what was happening at the bottom of the distribution than on what was happening at the top!

We spent a lot of time in our paper explaining what is wrong with this. Unlike ‘absolute poverty’, ‘relative poverty’ is said to exist when people’s living standards fall a long way short of what is ‘normal’ in their society. According to the Smith Family, this point is reached when people’s incomes fall below half the average income – and this is where they drew their ‘poverty line’. Measured in this way, one in eight Australians is ‘poor’, and poverty increased in the 1990s.

There is, however, a fundamental problem with measuring poverty in this way (and it is a problem which many researchers in this field readily acknowledge). If a small number of high earners increase their incomes, this will pull up the average income, thereby dragging the ‘poverty line’ up too. The result is that more people will now fall under this line, even though they may well have got better off. This is exactly what happened in Australia over the last ten years.

To avoid this problem, most researchers nowadays define the ‘poverty line’ as half the median (i.e. the middle-ranking) income, rather than half the average. Defined in this way, the number of Australians in ‘relative poverty’ comes out at 1 in 12 – not 1 in 8 as the Smith Family claimed. Furthermore, this proportion remained virtually unchanged throughout the 1990s.

Responding to all this, The Smith Family claimed that it calculated no fewer than twelve different poverty estimates, and that eleven of them showed that poverty had increased over the last decade. But this again really is misleading. Bear with us while we explain why.

Their original report actually identified six different ways of calculating a poverty line, each of which was repeated both before and after housing costs were taken into account. Because the Smith Family report chose to look at figures before housing costs, we followed suit. This leaves us with six estimates, not twelve.

Of these six measures, two were based on the well-known but badly-flawed ‘Henderson’ poverty line. Both of these produce absurdly-inflated estimates suggesting that getting on for a quarter of the population is in poverty! Wisely, the Smith Family rejected these measures, and so did we. That leaves four.

Of these four, one – based on half-average incomes – was the one they choose to use in their report, and it shows a rise in the number of people under the poverty line. The other three were all based on different versions of our preferred measure, based on half the median (middle-ranking) income. In all three of these cases, the number of people under the poverty line hardly changed over this ten year period (there was a slight increase according to two of these measures and a decrease in the third – essentially, the trend line is flat).

Claiming that 11 out of 12 measures all point in the same direction is, therefore, misleading. Some of these measures are effectively duplicates; others are useless; and of the four that are left, three indicate that nothing much changed over the last ten years.