Decision makers can be informed or mislead on policy decisions depending on the quality of supporting research. It is important to ask what the conclusions are, but also how those conclusions were arrived at in terms of assumptions, methods, and data reliability. A case in point is research on identifying a living wage rate for New Zealand.
Living Wage Aotearoa New Zealand commissioned King and Waldegrave to identify “the income necessary to provide workers and their families with the basic necessities of life. A living wage will enable workers to live with dignity and to participate as active citizens in society.” They calculated this to be $18.40 per hour, after government transfers are taken into account. [Comments on new rate at end.]
There are multiple issues with King and Waldegraves report.
1. The definition raises the question of what are “basic necessities” and what does the second sentence actually mean? King and Waldegrave made little attempt was to gain wider community input. “Basic necessities”, as used in my original report and by King and Waldegrave, includes items above (and including) what would be in a basic living costs budget. For example, a home computer may have community support as a basic necessity, but subscription TV not.
Treasury noted that only six per cent of low income households fit the authors’ two adult and two children household assumption. Only 21 per cent of all low income households have any children at all. The London Living Wage is based on a weighted average of eleven household types. Treasury concluded that a living wage is poorly targeted, and would adversely affect employment opportunities for young adults, in particular.
In addition, because of government transfer abatements and increased income tax, most of the additional income would be reclaimed by the government. King and Waldegrave state that a living wage could be adjusted up or down depending on changes to tax rates and government transfers. This raises an intriguing question: if a new living wage rate is less than the previous rate will actual wage rates and wages that rose in relativity, really decrease?
King and Waldegrave used alternative assumptions for the ages of dependents in calculating food costs, childcare costs, and government transfers. The effect was to inflate costs, and under-estimate government transfers. Additionally, childcare costs were based on ten hours of partially subsidised care at three dollars per hour (net), on top of the twenty fully subsidised hours. The hours and rate were not substantiated. Income from dependent’s part-time jobs were ignored, but the associated discretionary expenditure was included. Tax refunds for school donations and other refunds are not accounted for.
2. The second major issue is whether the methods appropriately measure basic necessities. King and Waldegrave used several methods to estimate household expenditure. Food costs were based on Otago University’s food cost research, and rent based on lower quartile rent data from MoBIE. Other expenses (e.g. clothing, communication, transport) were based on the average expenditure for income deciles 1 to 5 (1 is lowest income), for each category in Statistic NZ’s Household Expenditure Survey (HES). This last method results in a relative measure, not an absolute measure, of expenditure.
Costs for items that may not have public support as necessities were captured. For example, subscription TV packages, electronic video game systems, games of chance, pets, and overseas packaged holidays. Waldegrave has written subsequently that participating modestly in society includes overseas travel, but is his opinion representative of the wider community?
3. This method also leads to double counting of expenses: childcare costs were listed as a separate category in the report, but are also a cost component in the HES education category. Some items were inconsistent with the authors’ assumption that the household was renting: dwelling, and mortgage insurance. Other jurisdictions use methods that mitigate these issues.
4. The underlying HES has significant sampling errors. All surveys have random sampling errors, and Statistics NZ provides a sampling error figure for each value. Of the nine categories (based on the HES) used, at least five have been flagged by Statistics NZ with “care should be taken in interpreting these expenditure estimates” because of high sampling errors. These limitations should have been noted in the report.
The HES may also has non-sampling errors, and not a true reflection of the target population. Respondents may over-report what they spend on some categories (education, food) while under-report on others (alcohol). This could be due to social desirability, or because they perceive that the HES is used to set government transfers. Education spending ranged from $76.27 per week (lowest income decile 1) to $12.05 (decile 5). The figures do not make sense.
5. There is no discussion in the report that connects the $18.40 rate to relative increases in productivity or staff morale, although this is the stated reason Wellington City Council has adopted a living wage policy. Given the NZ report has different assumptions, a unique method, and data reliability issues, it is unsafe to assume the conclusions of overseas research are relevant in New Zealand’s economic, legal and social environment. What will an organisation do if productivity gains (and supposed other benefits) are less than required to offset increased wage costs?
Due to Equity Theory all other low income workers will want a pay increase, or will decrease output to compensate: the economic cost to firms, whether they pay a living wage or not, is equivalent to actually paying a living wage, but without any gains in overall output. [This has proven to be an accurate prediction: the Tertiary Education Union wants Massey Campus workers to be paid the same as the Warehouse Group’s living wage scheme, and the PSA wants relativity increases for staff earning above $18.40 at Wellington City Council].
It is astounding that local bodies such as Wellington and Auckland Councils are making multi-million dollar decisions based on the personal and subjective opinions of two individuals on what constitutes a living wage. Decision makers need to ask what the researchers’ conclusions are, and just as importantly, how they reached their conclusions.
UPDATE: New rate announced: $18.80. King and Waldegrave unilaterally changed the method and decided to add the increase in average wages (2.1%) to the $18.40 rate. The justification is that a living wage is now a wage in the market. This is absolutely contrary to their definition of a living wage.
They also calculate (using the original method) that the rate should be $22.89 per hour. Statistics NZ puts median wage in 2013 at $21.58 per hour. What method will they use next year? The big question is if this was the first year that they had calculated a living wage, would they have used the $22.89 figure?
To read Brian Scott’s full research paper:
A Review into the Basis for a Living Wage Rate in New Zealand
– click HERE
 Peter King and Charles Waldegrave. Report of an investigation into defining a living wage for New Zealand.
Family Centre Social Policy Research Unit. December 2012.
 Treasury Report: Analysis of the Proposed $18.40 Living Wage. 2013 September 30. 12 percent of low income households have two adults and one or two children. http://www.treasury.govt.nz/publications/informationreleases/livingwage
 Charles Waldegrave. The argument in support of a living wage. Dominion Post. 21 January 2014, p. A9.
 Pressure on uni over living wage. Talia Shadwell. Manawatu Standard. 10 May 2013. http://www.stuff.co.nz/manawatu-standard/8657335/Pressure-on-uni-over-living-wage
 Congratulations to Wellington City Council for stepping up to Living Wage challenge. PSA. 19 Dec 2013. http://www.psa.org.nz/newsroom/mediareleases/13-12-19/Congratulations_to_Wellington_City_Council_for_stepping_up_to_Living_Wage_challenge.aspx
 Living Wage Aotearoa New Zealand 2014 Update. King and Waldegrave. http://www.livingwagenz.org.nz/files/Living%20Wage%202014%20Report.pdf