Twenty years of radical economic reform and the stronger growth that followed have left real incomes for the poorest 30 per cent of New Zealanders lower than they were at the outset.
The gap between rich and poor has widened substantially from the egalitarian days of old; and the rate at which income inequality has grown is the fastest among developed countries.
These are among the revelations in a report on household incomes just released by the Ministry of Social Development.
It is based on Statistics NZ's household economic surveys, done face to face with a sample of 3000 households.
The surveys are now undertaken every three years. The most recent data available for this report was from 2004. Results of the 2007 one will not be available until next year.
As is normal for such studies incomes are adjusted for inflation and for differences in the size and composition of households to aid comparability.
Many of the report's findings confirm what one would suspect.
When it comes, to incomes it is best not to be old or young.
The poorest fifth, or quintile, of households (this is by income not wealth) includes one in every three people aged over 65 and one in every four aged less than 17.
That is before housing costs are taken into account. Looking at incomes after housing costs half of those over 65 are in the second lowest quintile and only 10 per cent in the lowest.
The converse is that 29 per cent of people aged between 45 and 64 are in the top income quintile (before housing costs).
Other factors which increase your chances of being in the top income quintile include ethnicity: 25 per cent of Pakeha are in the top quintile, but only 3 per cent of Pacific people and 10 per cent of Maori.
Another is childlessness: 41 per cent of under-65s in couples are in the top income quintile, but only 19 per cent if they have dependent children.
The children of a solo parent have a 57 per cent chance of being in a household in the lowest income quintile and only a one in a 100 shot of being in the top quintile.
So far so unsurprising.
But when the report turns to plotting how real incomes have changed over the 20 years from 1984 to 2004 at various points in the income distribution, the results are sobering.
If you ranked all the households in the country by income, after tax and inflation but before housing costs, the bottom 30 per cent in 2004 were worse off than the bottom 30 per cent in 1984 had been.
Even at the 40th per centile the gain was only 3.2 per cent altogether over the 20 years.
The top 10 per cent of households, by contrast, were at least 21 per cent better off in 2004, in terms of real disposable incomes, than their counterparts in 1984. The top 20 per cent were at least 16 per cent better off.
The median household, the one exactly half way down the list, was 9 per cent better off than its counterpart 20 years earlier.
The 20-year period was a game of two halves.
Between 1984 and 1994 incomes fell all along the distribution. That was the era of root and branch economic reform - liberating for many people but calamitous for many too. It included a global recession in the early 1990s as well.
In the 10 years after 1994 real incomes rose again, but more quickly for middle and high-income groups than at the bottom end who, as we have seen, did not fully recover to 1984 levels.
By 2004 the top 20 per cent of households were getting 40 per cent of the pie of national household income, five times the 8 per cent share of the bottom quintile.
That is very similar to the relative shares in Australia and the United Kingdom.
One of the standard measures of income inequality is called the Gini coefficient.
New Zealand's is higher than the OECD average, indicating more inequality. The rate at which it has increased since the mid-1980s is the highest among the 20 countries the report compares us with and more than three times that of Canada, Australia or the United States.
So how many people are poor in our increasingly unequal society?
It depends where you draw the line.
For example if you define "poor" as having a disposable income less than 60 per cent of the median, then one child in three is in a poor household.
If you define it as less than 50 per cent then the proportion of children counted as poor drops to one in seven.
On the former definition, the proportion of the population classified as poor has risen from 14 per cent in 1984 to 21 per cent 20 years later.
On the latter, 50 per cent of the median, definition the proportion has risen from 7 to 10 per cent over the same period.
An alternative definition of poverty is the "constant value" approach which takes some measure of poverty and adjusts it for inflation. This gives a sense of whether the incomes of low-income families are rising or falling in real terms.
"Whatever is happening to the non-poor, if more and more people end up falling below a constant value threshold, as happened in New Zealand from the late 1980s to mid-1990s, then there is likely to be wide concern about increasing poverty," the report says.
On that basis, taking 60 per cent of the the 1998 median income as the benchmark, the proportion of the population classified as poor doubled between 1984 and 1994, from 13 to 26 per cent, before declining to 13 per cent again by 2004.
Who are they?
The majority, 64 per cent, are households with children.
One positive trend is that the proportion of households made up of a sole parent with children falling below that poverty line has declined from 72 per cent in 1996 to 55 per cent in 2004.
Likewise the proportion of households made up of people under 65 whose main source of income is a benefit who fall below the poverty line declined from 66 per cent in 1994 to 56 per cent 10 years later.
But ethnicity remains a factor.
Maori children are nearly twice as likely as Pakeha children to fall below the poverty line and "others" (including Pacific people) nearly three times as likely.
Even with a tight labour market the proportion of children in "workless" households in 2004 was 14 per cent, down from 22 per cent in 1994.
The Working for Families package rolled out over the past three years will put an additional $1.6 billion this year mainly into low to middle income families.
It will be interesting to see how much difference that makes to these dispiriting numbers.