Treasurer should put his Ginis back in the bottle and go for growth

THE genie is out of the bottle and no statistic is safe. For clutching his Gini coefficients, Wayne Swan is on the rampage. The Treasurer's fury is directed at this paper's George Megalogenis, who writing in last Thursday's edition dared suggest inequality had increased on the class warrior's watch.

Not so, bellowed the great man, unleashing a whirlwind of press releases, tweets and twitters that called Megalogenis's claim "clearly incorrect and (based on) no evidence."

So let's start from the beginning. The Australian Bureau of Statistics' most comprehensive attempt at measuring the distribution of income, taking account of the taxes people pay, the social benefits they receive and the public services they consume, is its series on Government Benefits, Taxes and Household Income.

Because of the complexities involved in producing these estimates, they are not available for every year, but a comparison can be made of a year in the Howard government, 2003-04, and a year under Labor, 2009-10.

What that comparison shows is that the income shares of the lowest income groups were all lower in 2009-10 than in 2003-04, while the income share of the top quintile was higher.

Bear in mind that measure takes account of all taxes paid, benefits received and public services consumed. And there's an additional important point to bear in mind. The ABS, bless its soul, provides the information needed to assess whether the observed changes are statistically significant: in other words, whether one can be reasonably confident that they are not mere vagaries of data collection.

And in this case, yes, they are significant and substantially so (all the details are on my blog). As a result, one can be pretty sure the distribution of income was somewhat less equal in 2009-10 than in 2003-04.

As for the great man, this is what he said in rebuttal: "The Gini coefficient for equivalised disposable household income - a measure of income inequality that increases as inequality does - rose from 0.306 in 2003-04 to 0.336 in 2007-08 under Howard but has since moderated to 0.328 in 2009-10." But does that "conclusively point to decreases in inequality under the current Labor government", as Swan went on to claim?

Far from it. For here, so to speak, is the rub: the change is not statistically significant. As a result, one cannot infer that the Gini coefficient has actually changed at all since Labor came to office, much less state "conclusively" that it has declined, despite all the cash Labor has splashed around.

Of course, it may be that Swan has no idea what statistical significance means - after all, you could hardly learn that from Bruce Springsteen songs. But The Age's Tim Colebatch, who yesterday repeated Swan's claims, should know better. And far from this paper's claims being "false", as Colebatch states, they reflect what we can say with confidence: which is that incomes were somewhat more evenly distributed in 2003-04 than in the most recent year for which we have reliable evidence.

But there is a much bigger and more important question, which is this: why is the Treasurer so fixated on income distribution? After all, experience and analysis point to a simple fact: the best way to make everyone better off is to grow the pie, not to waste resources robbing Peter so as to pay Paul.

And that is unsurprising, as economic growth is not a zero-sum race, in which only the winner gets gold: rather, it creates opportunities for all.

Indeed, the ABS data makes that clear, for while the income share of the two bottom quintiles fell by about 1 per cent between 2003-04 and 2009-10, strong economic growth ensured their earned incomes increased by more than 20 per cent in real terms. No plausible scheme of redistribution could have secured so great an increase, and in Australia's history none has.

Swan should therefore put his Ginis back in the bottle. And once he's done so, read a good book on statistics. Not to mention economic growth.