Max Rashbrooke asserts that the “most fundamental omission” in our report is “its failure to deal in any significant way with the long-run consequences of widened inequality”.
Elsewhere he comments that “The biggest issue I have is The New Zealand Initiative is completely oblivious to the point that even if the big increase in inequality was in the 1980s or 1990s, and hasn’t worsened since then, it still has big implications for the country today”.
Note that Rashbrooke’s comment does not contest point 8 in our 21 Messages blog–it is wrong to assert that disposable income inequality IS rising in New Zealand or even to assert that it has risen since the 1990s, at least on the most cited measure. However, his “even if” framing suggests that he refuses to accept the evidence. It would help public debate if he clarified his position on point 8.
Rashbrooke’s fear/assertion is that the long-run consequences of the earlier rise are “incredibly damaging”. In his view, inequality per se “has strong and internationally accepted impacts on things like social mobility, trust, health and social cohesion”.
Next he immediately conflates income inequality with poverty or hardship by asserting that “[j]ust because poverty isn’t skyrocketing at the moment does not mean there isn’t a problem to deal with”. Of course it doesn’t, but what has that got to do with income inequality?
Our first point in response is that on the evidence there has been a long-term decline in income inequality, at least from the 1950s. Figure 3 in our report reproduces income share calculations by Atkinson and Leigh. It shows a long decline in income inequality, as indicated by the share of the top income earners, at least from the 1950s.
But there is a sharp blip upwards in the late 1980s and the early 1990s.
The key question is the robustness of that measured blip. Did CEOs and other top earners really manage to increase their share of pre-tax private income to a major degree amidst the aftermath of a major sharemarket and commercial property crash and around the height of New Zealand’s biggest recession since the 1930s ? Most people would look for other explanations.
As we note on page 17 of our report, Brian Easton has independently reworked the figures behind out Figure 3 to eliminate the effects of the introduction of dividend imputation in the last 1980s. Below we reproduce, with his permission, one of his charts from a presentation he gave in 2014.
These calculations indicate that the upwards blip was an artefact of the introduction of dividend imputation. Pre-tax personal incomes at the top end rose in the 1988-1990 tax year for a technical reason. An unchanged cash dividend received by a shareholder was grossed up by the amount of the dividend imputation credit that was now attached to it. So personal pre-tax income rose for an unchanged dividend cheque.
On these figures there has been greater top end pre-tax market income equality in New Zealand since 1990 now than in any time in the 1950s, 1960s or 1970s. If there really is a strong connection between pre-tax market income inequality and social outcomes, Rashbrooke should be able to document “incredibly beneficial” consequences.
What about post-tax market income inequality? Certainly, the elimination of the double taxation of company income will have lifted top after-tax market incomes, other things being equal, because less tax was due at the personal level on unchanged cash dividends. So will the major reductions in the top rates of personal income tax. But other things were not equal. Section 1.4 shows that personal income tax paid collectively those paying the top rates of personal income tax increased markedly during this period. More research is needed, but the government’s tax base broadening measures and the alignment of tax rates were surely playing a role.
A second point to note is that Rashbrooke does not dispute the Ball and Creedy finding that spending was as equal in 2014 as in the mid-1980s. (See points 6 and 7 in our 21 Messages blog and our blog here on the relevance of spending inequality evidence.) Can this outcome really be consistent with “incredibly damaging” long-term consequences?
A third point is that Rashbrooke’s defence of The Spirit Level against expert criticisms fails to address any matters of substance. Instead he confuses the pre-publication screening by two or three referees of an article submitted to a refereed journal with the post-publication assessment by readers with relevant expertise. The former review merely answers the ‘fit for publication’ question. In no way is it an opinion about whether the article will come to be regarded as authoritative.
A major criticism of The Spirit Level is that correlation is not causation. Suppose a full-time productive worker has a stroke or heart attack and can’t work again. Sure enough, the statistics will show that lower health quality was associated with lower income. But the authors of The Spirit Level’s obviously cannot prove that the stroke or heart attack, accident, sickness or whatever was caused by the low income that was its consequence. Arm-waving about a single peer-reviewed article does not address this criticism.
We list peer criticisms of The Spirit Level in footnote 7 in our report. Eric Crampton’s blog post here draws particular attention to those by former ANU Professor of Economics and current Labour Member of Parliament in Australia, Andrew Leigh. Leigh makes it clear that he wanted to believe Wilkinson and Pickett’s claims, but found they fell apart under scrutiny.
In the same comment Rashbrooke also asserts that we fail “to deal with important issues such as discrimination, politics and power”. In fact, in chapter 5 we emphasise that it is important that incomes and wealth are fairly and legitimately earned–in fact and in perception. We caution against corporate welfare for this reason. See also our concluding chapter, the last point in our Key Findings and point 18 in our 21 Messages.
Also it would be good to see actual evidence that unequal outcomes are due to market power. For example, how could market power explain the increased tax take from top income earners that we document in section 1.4?
Point 9 of our 21 Messages refers to marked inequalities in paid hours worked and hourly earnings by educational qualifications and age (as a proxy for experience and responsibility). It is not clear how Rashbrooke has adjusted for these aspects when making strong assertions about the contribution of male/female discrimination.
The bottom line is that Rashbrooke’s fears about adverse long-run consequences from increased inequality don’t apply when the long-run trend is downwards or flat. It does not help to cherry pick the time periods and the inequality measure. The problem of hardship in too many households confronts the community regardless of whether economic inequality is rising or falling. Rising house prices and costs are a stand-out issue, but not the only one.
In a nutshell, we do not see that Rashbrooke’s criticisms of our report impinge materially on our 21 Messages.
The Sand Pit 