What they're saying about us in Luxembourg

I reported last month on a simple exercise I’d done that seemed to show the post-tax distribution of income in Australia becoming more unequal under the Howard Government. In response, a few readers wrote that these figures seemed to be at odds with data from the Australian Bureau of Statistics, which showed a drop in inequality from 1996-2003. To confuse us all further, the Luxembourg Income Study has just concluded an agreement with the ABS to add Australia to its inequality comparison.* Here’s what they find:


Gini Atkinson (e=0.5) Atkinson (e=1.0) 90/10 90/50 80/20
1981 0.281 0.066 0.140 3.93 1.86 2.48
1985 0.292 0.072 0.149 3.97 1.87 2.55
1989 0.304 0.077 0.157 4.19 1.94 2.59
1995 0.308 0.081 0.171 4.02 1.94 2.70
2001 0.317 0.085 0.176 4.25 1.99 2.73
2003 0.312 0.082 0.172 4.24 1.98 2.71

It looks to me as though LIS thinks inequality has indeed risen under Howard, though by less than it did under Hawke and Keating.

* For the data nerds, LIS is the best data for comparing income distribution across countries over the last couple of decades.

This entry was posted in Inequality. Bookmark the permalink.

8 Responses to What they're saying about us in Luxembourg

  1. Andrew – What’s driving the slight drop 2001-2003?

  2. Andrew Leigh says:

    I’m not sure. Rising family benefits, perhaps?

  3. Bruce Bradbury says:

    Unfortunately there are quite a few continuity breaks where the (ABS) survey methods changed. The biggest break is between 1989 and 1995, but there might also be a break before the last year.

    LIS haven’t put up documentation yet, but I’m guessing the last year result is based on current weekly income for the 2003-04 year. One feature here is that the May 2004 budget provided a lump-sum payment of $600 per child (paid in June 2004). The ABS have imputed this as a payment to all families in the survey in the 2003-04 period. That is, weekly income is increased by $600/52 per child.

    There are other odd features of the 2003-04 data that I am currently looking at. In the household surveys, real median equivalent household income jumped by about 6% between 2002-03 and 2003-04 (compared to about 2% for mean income in the National Accounts). About a quarter of the jump is due to improved measurement of self-employment income. I have yet to calculate how much is due to the family payment mentioned above.

    Peter Saunders and I have just published a paper in the Economic Record where we look at poverty trends between 1993-94 and 2002-03. For relative poverty (the fraction of people below half median equivalent income) we conclude that

    For overall relative poverty, these imply a modest increase in poverty during the middle of observation period that has flattened out since 2000–01. For child poverty, the overall pattern is more U-shaped, with poverty rates slightly higher at the end of the period than at the beginning.

    (‘modest’ here is 1-2 percentage points). Taking the 2003-04 data at face value, they suggest that child poverty has fallen by about 1 percentage point and that overall poverty has levelled off. This is consistent with the family benefits story, but I havent’ checked it out properly.

  4. Peter Whiteford says:

    And you get a slightly different story again from the OECD Income Distribution study, which can be found at http://www.oecd.org/dataoecd/48/9/34483698.pdf

    The easiest summary is in Table 1.

    This is not quite so recent as the LIS data (goes up to 1998-99) and it is based on the Household Expenditure Surveys for Australia rather than the Income surveys (the numbers as Bruce knows are produced by colleagues of his at the SPRC) and use a common methodology across countries. The OECD data overall have fewer counries than LIS but we do have Japan and New Zealand and until the recent LIS release they were more up-to-date on Australia.

  5. Christine says:

    The increase in inequality over the 80s didn’t occur as much in other countries, isn’t it?

  6. backroom girl says:

    All I can say is that if the amount of extra money the Howard government has thrown at families (especially single parent families) hasn’t cured child poverty, it seems to have proven fairly conclusively that extra money through family payments is not the answer.

    The increases in real value of family tax benefit have been so large under this government that, when combined with the indexing of single parent pensions to increases in wages, I would find it literally unbelievable that single parent families were just as (or more!) likely to be ‘poor’ than ten years ago.

  7. Peter Whiteford says:

    Backroom girl’s point about child poverty is very interesting. A colleague and I will have a new OECD working paper about this out on our website in the next week or two. When you compare benefit entitlements in 1999 to the poverty line – i.e. what people should be getting compared to what they say in the survey, then indeed benefit levels in Australia are at or above the poverty line (50% of median income). There may be a problem of take-up of benefits, and the results also depend on how you adjust for family size. But there is the problem of understatement of incomes at the bottom of the distribution that the ABS has acknowledged. What would be useful is for the ABS to go back to people after they have interviewed them and ask them why they say they are getting less than their entitlements.

  8. Leo says:

    This is all ‘relative’ poverty isn’t it? Nice measure of inequality, but has anyone done any reviews of ‘material’ poverty since Howard began ratcheting up family payments? Peter Saunders (the CIS one) had some interesting numbers on ‘hardship’ in his book “Australia’s Welfare Habit”. It would be nice to know if the steady increases in family payments are having any impact on that front.

Comments are closed.