Truth is a strong word when discussing inequality …

By Matt Nolan 24/01/2014


Over at Polity Rob Salmond has promised us the truth about the gap between NZ’s rich and the rest (via Toby Manhire)!  This would be encouraging, given the complexity of the data.

I just want to say at the start I have full respect for people who want to discuss these important issues, including the use of data.  However, after reading his post I think he may have oversold his claim.  Don’t get me wrong, what he posted was interesting – if you go over to his post you will see a graph that shows aggregate taxable income for three groups – those being taxed on $150k a year, those being taxed on $100-$150k a year, and those being taxed on less than $100k.

I do not have the data sadly, but I have some reservations stemming from what I see in the post.  For some reason he is only quoted gains since 2010/11 – ignoring the whitewash for high income earners that occurred in 2009/10 due to the global financial crisis.  Furthermore, the income changes he quotes are biased (to the point that they aren’t representative of household income at all) in two ways:

  1. These are not per person figures – people have been moving “up” income bands, so the proportion of the population in the highest band will have gone up.  As a result, this exaggerates income growth. [Note:  I’ve been informed I didn’t point out how significant this is – the movement in groups in the income bands is likely to be enough to explain ALL of the change, making the figures quoted in the Polity post virtually incompatible with any clear narrative for talking about per capita income.  I failed to point this out sufficiently as I thought it was self evident – but that was a stupid assumption on my part, sorry.]
  2. This ignores the explicit push since 2010/11 to improve taxation on some types of capital gains and catch tax evasion.  This will have increased estimates of “taxable income” especially at the top end – although this does imply that previous figures for the “top end” will have previously been understated (a common issue).
  3. I am not sure if the claims here are about individual or household incomes – however, we need to be careful comparing it with any “base” without defining this.

If we are interested in these ideas, why don’t we look at reports based on the Household Economic Survey data – which specifically looks at the full distribution of household and personal incomes?  Or we could even look at the decile data in the HES directly, and compare what has happened to the top 10% relative to other groups!  Why don’t we also have a little peek at the research about what has been going on?

The “truth” of the issue is that this is very complicated.  We have seen median income growth outstrip mean income growth in NZ for a long period of time, implying that static inequality has come down a little bit.  But a single inequality index cannot tell us whether what is going on is fair or just, instead we need to ask harder questions about poverty, capability, and opportunity within society.

A misleading graph (sorry but it is, as I’ve explained) titled under truth doesn’t do this.

And before you pin me down as someone on the right you ideologically shouldn’t listen too, remember that I have written in support of thinking in terms of relative income poverty, and stated that we should have a minimum income based on a land tax.  But I also realise that New Zealand is not like the United States, and trying to squeeze our data into their narrative (instead of discovering our own) will lead to bad policy.

Update:  A friend of mine pulled this graph together, showing growth in the population of people in the differing tax bands – in order to illustrate the first point I made above:

image002

This shows that the vast majority of the difference in income growth in Rob’s post is due to having more people in higher bands, not due to actual income growth for individuals/families earning high incomes.

This is why looking at the HES data tells such a different story – and for the narrative Rob was painting (discussing the income of individuals and households, with some doing well and some poorly) the HES data is the correct data, while the aggregate tax data is wrong (due to population changes in the bands).  It implies that saying high earners have been “creaming it” is factually incorrect – not as a matter of opinion, but in actual data.

We can argue about what is fair, and what is just, of course.  But let us also be careful with our data!