Friday, August 19, 2005

A picture is worth a thousand words

While digging for information for this post, I came across the following graph in a Treasury report which makes it abundently clear who were the winners and who were the losers from the Revolution. I've covered this ground before, but this one graph tells the story better than I have ever been able to do:

The graph is in real (1998) dollars, and it clearly shows that 70% of households lost out in real terms from the reform process - they were worse off when they started. Meanwhile, the top 10% made out like bandits, increasing their income by 30% while everyone else's declined or stayed stagnant.

The impact on poorer families of National's 1991 benefit cuts is particularly noticable, especially when compared with the skyrocketing incomes at the other end of the graph.


correct me if I'm wrong, but you seem to make a base assumption that smaller income disparity = greater social justice. i don;t believe it does.
Social mobility, ie the ability to increase one's social 'class' as a result of effort, seems to be truly motivating to people. The trick is to allow people to better themselves without treading on others.
In that sense, you can't automatically conclude that increasing income disparity is bad.. the tall poppy syndrome was an incredibly stifling factor to individuals excelling.

Posted by Anonymous : 8/19/2005 01:29:00 PM

Egalitarianism is at the heart of the kiwi dream. But the really telling point in the above isn't about a relative decline in people's incomes (in that those of the rich grew faster); it's about an absolute one. The reforms actually made people significantly worse off, and those on the bottom of the heap suffered the most. In effect, they robbed from the poor to give to the rich - and that doesn't fit with any conception of social justice.

Posted by Idiot/Savant : 8/19/2005 02:00:00 PM

The problem with divisions into deciles is that decile 10 is the only decile that has an unbounded upper income thus is exceptionally affected by outliers. That is, a few very rich people will pull its average way up compared to the effect the same number of people would have on lower deciles. You can see it in the graph as decile 10 is the only one that doesn't show a dip in the middle, whereas decile 9 and below do to a greater or lesser extent. I'm surprised this isn't mentioned in the report since it is a well known problem in such studies. There are better ways to show what they purport to show.

I also don't see how the "1991 benefit cuts" are visible. 1991 is where every decile bottoms out and turns upwards.

Posted by Chefen : 8/19/2005 06:19:00 PM

Actually 1991 is where the upper deciles took big hits. Go figure.

Posted by Chefen : 8/19/2005 06:28:00 PM

Sorry to comment again in a row, but I can't see what you are getting at. There is a definite downward trend through the 80s, but what was it prior to that during the Muldoon years? I find it hard to believe that people were getting better off then. The numbers for those years would be nice. Is the downward slope better or worse than the years before? Because that has an enormous bearing on the conclusions you draw about the "reforms". But most of all, *what* caused the turn around starting at about 1991?

Posted by Chefen : 8/19/2005 07:58:00 PM

This is an indication of share of govt income and low tax rates for low income earners and high marginal taxes for those at the top affect incomes (as it was in 1980) to a point of medium level taxes for low income and medium taxes above that.

Posted by Anonymous : 8/19/2005 09:29:00 PM

The more I think about this, the more I doubt your conclusion - there's no demonstrated causality. Other social factors have had a huge influence over the past decades.
The first to mind and perhaps the biggest is the change in the nature of work. Technology and globalisation have had massive impacts on the types of work undertaken. At the lower end, manufacturing has been replaced by mechanisation and/or outsourced to cheaper labour countries, reducing the pool of skilled manufacturing jobs. Many of those have been replaced by commodotised service-sector jobs. The unions have been largely broken, and ongoing company aquisitions and relentless driving down of costs have reduced pay rates in service sector jobs.
At the top end, technology workers are in high demand and short(er) supply, and salaries have increased to reflect this.
This is certainly a revolution, but not necessarily the one you're implying.
Of course the reforms of the 80's have had huge impacts, and the ridiculously naive faith placed in market forces meant the most vulnerable were least protected, but I suspect the causes of the changes are as much global 'restructuring' as they are of local.
It would be interesting to compare NZ changes with other countries to see how they aligned (or not).

Posted by Anonymous : 8/19/2005 10:14:00 PM

Tane, according to the report Appendix the deciles are composed of equal numbers of individuals. If it were by income then the logical thing to investigate would be the number of people in each income band... a more meaningful measure for what is being discussed here. Decile 10 is defined as the top 10% of income earners regardless of their incomes, so there is no real limit. The other deciles are bounded by the bottom of the decile above. Averaging doesn't get rid of the problem, the big earners at the far end of the distibution skew it strongly for decile 10, whereas they don't in the other deciles unless you have a strange (eg bimodal) distribution over all incomes.

The graph shows something, just not what is being implied here and the lack of "pre-reform" data is a bit suss in my mind.

Posted by Chefen : 8/19/2005 10:41:00 PM

Here is a link to a recent
paper that updates the trend to 2004 - see Figure 1. It's a different measure - quintile medians rather than decile averages, but the trend is the same.

For an international comparison of inequality, see The Social Report 2005:

"Comparisons with other OECD countries are available using a different measure, the Gini coefficient.54 Gini coefficients measure income inequality, with a score of 100 indicating perfect inequality and a score of 0 indicating perfect equality. Around the year 2000, New Zealand’s score of 33.9 indicated higher inequality than the OECD median (30.1) and a ranking of 18th out of 25 countries. Northern European countries had the least income inequality, Denmark ranking lowest with a Gini coefficient of 22.5. New Zealand’s score was slightly higher than Canada (30.1), Australia (30.5) and the United Kingdom (32.6), and lower than the United States (35.7).55 The 2004 figure for New Zealand was 33.5."

Posted by Anonymous : 8/19/2005 10:48:00 PM

It might also be worth noticing some of the conclusions in the report

The key finding which this study reaches, however, is that government intervention—
through taxes and social expenditure—has maintained the incomes of less well-off
households. While market incomes in lower deciles declined, net government benefits
have increased by at least the same amount. Between 1987/88 and 1997/98,
redistribution became more favourable toward households in the low income deciles. A
part of this increase in redistribution is automatic—households earning more income pay
more tax, for example, but are unlikely to receive more in government benefits—but a part
of the increase is also likely to be a result of specific government policy changes. As a
result of this change in redistribution, real incomes were maintained for all income deciles
between 1987/88 and 1997/98. While there will undoubtedly have been individual
households which suffered financial hardship and declining circumstances, all deciles
were, on average, doing at least as well in 1997/98 as they were ten years previously, and
in most cases were doing better.

It is also clear why there was a bounce upwards from 1991/92, Figure 1 shows the real GDP going from flatlining to increasing at the same time. So with a slight lag, the incomes are following this, over all deciles. Notably it turns down again in 1997, so if the income data was continued I'd expect you'd see a levelling off of incomes after 97.

The report is for 87-97 which is why no previous data is there. The economy was clearly in the shitter prior to that with the downward trend in GDP and incomes. Something caused a turn around in the years leading up to 1991/2.

Posted by Chefen : 8/19/2005 11:37:00 PM

Chefen: the average of the top decile is almost certainly distorted by outliers (they have enough of an effect on the average compared to the median). But what this means is that instead of 70% of us being fucked over for the benefit of the top 10%, we have 70% of us being fucked over for the benefit of the top 1%. Is that somehow supposed to be better?

Posted by Idiot/Savant : 8/20/2005 02:49:00 AM

I would expect most politicians (including labour) would probably say the country had dug itself a hole and that it is not surprising there was a drop for most of the country. Having said tht I think personally it was a little too much ideology and not enough pragmatism.

Posted by Genius : 8/20/2005 07:51:00 PM

I dunno I/S. The report shows that all deciles increased 87-97 after downward trends on all deciles except 10. You can try and nail the 1% at the very top who will manage to get better off whatever happens, until the economy crashes into the ground, but what you were claiming isn't shown by the graph. Was all I was trying to say.

Posted by Chefen : 8/21/2005 01:58:00 AM

Chefen, the most obvious thing to me about this graph is that at the end of the period, deciles 1-7 were making _less_ money than at the beginning, deciles 8-9 had just managed to crawl back, and decile 10 were making out like bandits throughout most of it.

And that's what Idiot said. Are you trying to say that the graph *doesn't* say this? I'm not saying it doesn't say other things *as well*, but there's no denying the drop in income for 70% of NZ.

Posted by Anonymous : 8/21/2005 04:38:00 AM

It's like arguing with a brick wall.

The graph is in real (1998) dollars, and it clearly shows that 70% of households lost out in real terms from the reform process - they were worse off when they started.

No, it doesn't. It doesn't show the trends prior to the reforms so you can't draw any conclusion, it may well have arrested an ever steeper decline, for all the graph shows.

Meanwhile, the top 10% made out like bandits, increasing their income by 30% while everyone else's declined or stayed stagnant.

Again, all deciles were in decline which then turned around about 1991. As the report says, all deciles were better off 1997 than 1987.

The impact on poorer families of National's 1991 benefit cuts is particularly noticable,

No it's not, look at the graph. Everyone did better from 1991, except decile 10 took a hit around 1991. So?

What it really shows is everyone in decline until around 1991 and then everyone turned around. The top most gained at a faster rate, which is pretty much what you'd expect since having more money allows you to create more. It doesn't show the mobility of people, some of those 1997 decile 10s could well be 1990 decile 3s... unless you believe that everyone stuck to some predefined decile and that class mobility is a fiction.

Posted by Chefen : 8/22/2005 06:36:00 AM

Chefen appears to be deliberately missing the overall point.

OK, try this one... note the last 4 years and see how much each decile has increased. On my old crappy monitor Decile 1 has not changed, 2 the change is only slight, 3 not much better, etc. Compare those to the drastic upward numbers for Decile 10.

Now the 10s pay the same as the rest of us for a kilo of apples and a kilowatt of power. Politicians of all colours (apart from ACT) say they don't mind paying more tax as they can afford it.

Why do I get the impression that the top 10th of our earners begrudge sharing it about a bit, especially on the number one priority which is (or SHOULD be) our children?

Maybe the brick wall chefen is sick of hitting is called Egalitarianism.

Posted by Anonymous : 8/23/2005 07:16:00 PM