Colin James, Statistics Forum, Te Papa Tongarewa, 25 March 2010
As a journalist, I long ago learnt that statistics are not what they seem. After I started using statistics, numbers lost the authority they had when I first counted to 100. Nevertheless, they have the patina of accuracy, which is of great use in journalism, either as shorthand or as convenient evidence.
Examples: global warming numbers; New Zealand’s ranking in countries’ ratio of prisoners to population; whether this society is becoming more or less unequal; ACC finances.
Politics, which is my beat, abounds with promiscuous statistics: the same numbers “prove” opposing policies or perspectives. In the 2002 election campaign the public and politicians were bamboozled by the difference between probability and proportion in the great corn genetic modification beat-up, which cost Labour several percentage points. Politicians are prisoners of opinion polls because they have acquired a statistics-like patina of accuracy. They use polls to develop and test policy, which is sad, because polls are backward-looking snapshots, not forward-looking oracles.
Some official statistics are polls — the household labour force survey produces unemployment numbers that are multiples of the actual numbers on the unemployment benefit. Economic forecasters pay close attention to that survey and to consumer and business confidence polls. Some of the questions in the census seem to me to verge on opinion polling. Ethnicity seems to me to be one of those.
Imagine how I explain to an intelligent international readership that statistically the population of this country is 110% of those who live here. This can only be because there are fewer really in each group than the statistics pretend there are — double-counting has a special magic. How do I explain that actually more people have Maori ancestry (18%) than are “ethnically” “Maori” (15%)? Given that Statistics New Zealand officially states ancestry (which I take to be a polite word for “race”) is a marker of ethnicity, this unsettles my commonsense.
So I treat those numbers as indicative. To say 15% of the population is Maori and 7% Pasifika and the percentages are much higher among the under-25s is to say something useful about the workforce’s present and future composition and thereby, if ethnic characteristics have a bearing on average productivity, to pose some questions for business investors. To say 3% identified as Asian in the 1991 census and 9% in 2006 depicts an interesting trajectory, even if a large number of those are transitory students: it says something about our geography and our economic and cultural futures.
But the numbers are not collected for my amusement or rough-and-ready, rule-of-thumb journalism and musing. They are collected, according to Statistics New Zealand’s official ethnicity report last October, as a basis for “stakeholders” to make policy and deliver government programmes. “Ethnicity statistics are used nationally, regionally and by communities to help identify demand for public policies, programmes and services, for tailoring their delivery and for monitoring the results,” the report said. It gave as examples, the health sector (funding, tailoring programmes and monitoring results), local government (particularly those “experiencing significant demographic, social and economic change”) and “Maori and other ethnic community group service providers” who “rely on official ethnicity statistics in planning and engaging with government agencies”.(1)
This appears to reflect a presumption that the numbers are helpful in enabling the government and other agencies to reduce disadvantage. This is a worthy objective, on ethical, social cohesion and economic performance grounds. But are the numbers actually helpful?
David Bromell in his 2008 study says this: “‘Maori’, ‘New Zealand Europeans’, ‘Pacific peoples’ and ‘Asians’ are reported on as if these are discrete, stable and more or less homogeneous social groups. This does not provide a particularly accurate or enlightening picture of relative advantage and disadvantage within New Zealand society.” I draw two inferences from that comment: the numbers are rubbery; and other measures of disadvantage may be more appropriate.(2)
Take the second inference first. Bromell directs attention to Simon Chapple’s work in 2000, which (in Bromell’s words) “marshalled evidence intended to show that disadvantage in New Zealand is more closely tied to age, marital status, level of literacy, educational qualifications, skills and geographic location than it is to ethnicity, broadly conceived”.(3) Well, yes, except that there do appear to be disadvantages associated with culture or ethnicity which compound other disadvantages: a person may be poorly educated, in poor health, etc, in part because they share certain personal, family, neighbourhood or cultural — that is, ethnic — characteristics and not just by coincidence. So there is a point in an ethnicity check. But do authorities treat ethnicity as deterministic and as the primary factor?
I ask that because the second inference I draw from Bromell’s comment is that multiple counting may overstate the ethnic factor in disadvantage. People are double-counted and are likely to be more so as people mix up and as Maori ancestry drops for some to 1/128th and 1/256th. Agencies apparently have mechanisms for discounting that (up to a point). But it might sharpen the discounting if Statistics New Zealand asked two questions: Which ethnic group do you most identify with? Do you also identify with another group or groups?
The ethnic factor may also overstate disadvantage because each ethnic group contains advantaged people. Note in this context that only 56% of those who declared Maori descent chose the Maori roll in 2006. Some at least of the 44% who chose the general roll are not disadvantaged; some National MPs say they know from conversations that many Maori vote for them in their general electorates. Other Maori have joined Ngati Kanguru, that is, they have migrated to Australia. They do this for a variety of reasons, including to make more money, but among them is to escape the negative underclass stigma assigned to Maori in pub-talk and in the media, a stigma underlined and encouraged by the wide use and misuse of ethnicity statistics. And I have come across people who say they are not Maori but are on the iwi roll, as if it were membership of a club. Who would not be, if there is a dividend?
That brings me to a broader point. Who are the “stakeholders'” target? Predominantly Maori and Pasifika. And of those two, Maori are more numerous and they have the Treaty of Waitangi. There is a notion that the Treaty obliges the state to atone for the damage of colonisation (and there was damage), which includes, under the authority of article 2’s tino rangatiratanga guarantee, special, including Maori-run, mechanisms to reduce disadvantage.
Actually, addressing social, health, educational and personal disadvantage is an article 3 matter, a matter of citizenship. The modern concept of citizenship connotes full participation in society and the economy, which implies state-guaranteed action to reduce inequalities of opportunity. Moreover, the modern concept of that action includes ensuring that assistance works. And that implies sensitivity to cultural and other differences, which in turn includes understanding and working with different worldviews. That is the genesis of and mandate for Tariana Turia’s whanau ora project and Maori entities delivering education, health and social services. That is not an article 2 matter, dealing with iwi and hapu autonomy. It is article 3, dealing with citizenship.
Are ethnicity statistics needed for this? Or is it a matter of what works and who can make it work? Is it a matter of Maori (and Pasifika) self-selecting into programmes which are more sensitive and applicable and for iwi or urban marae or other organisations (if necessary, with state funding) being available and facilitating and encouraging that self-selection?
Paul Callister has argued that in any case “by early 2007 ethnic-based special measures formed only a small part of the policy tools being used to try to overcome ethnic disadvantage.” If he is right, the reason given in the October report for collecting the statistics — that “stakeholders” find them useful or necessary — seems to lose some force. (4)
Which leads me to the issue the October report was intended to address: the large increase in 2006 in those who wrote that their ethnicity was “New Zealander”. Statistics New Zealand hopes an information campaign might reverse this refusal of racial categorisation and/or assertion of national identity. Maybe. But there is a growing sense of what it is to be a New Zealander and the ancestry lines are growing longer and ancestry is a basis for ethnicity. This is the opposite of the “invisibility” Tahu Kukutai assumed. It is an issue of indigeneity, on which I have written and spoken about elsewhere. [I think I will put down Aotearoan next year.]
In fact, the “New Zealand European” and write-in “New Zealander” categories could be designated the “no-special-treatment” ethnic group, since that appears to be the obverse of the point the “stakeholders” made to Statistics New Zealand about the other groups. The worry for “stakeholders” in the “New Zealander” choice is that target ethnicities might be under-counted because, the October report says, “the inclusion of a ‘New Zealander’ tick-box in the census ethnicity question would cause a significant number of people to report a sole ‘New Zealander’ response, who would have otherwise reported a ‘non-European’ response, with or without a ‘New Zealand European’ response.”(5) But what’s wrong with that, if it reflects a maturing nation and if the rambling list of ethnicity criteria includes common culture, common language, common ancestry and common geographic origin, all of which a “New Zealander” can claim as indicative of a group he/she belongs to?
Well, what’s wrong with it is this: journalists wouldn’t have lovely rubbery numbers to make a story out of; journalists wouldn’t be able to stigmatise ethnic groups with accurate-sounding negative statistics, thereby encouraging them to escape to Australia; journalists’ employers wouldn’t be able to make as much money. Please keep the ethnic statistics just as they are. ————————————–
1. Statistics New Zealand, Final Report of a Review of the Official Ethnicity Standard 2009 (Statistics New Zealand, October 2009), p1
2. David Bromell, Ethnicity, Identity and Public Policy (Institute of Policy Studies, 2008), p297
3. Bromell, p297
4. Paul Callister, Special Measures to Reduce Ethnic Disadvantage in New Zealand (Institute of Policy Studies, 2007), p101
5. Statistics New Zealand, Final Report, p19