Measuring the pay gaps

CEVEP monitors the pay gaps by gender and ethnicity using average hourly ordinary time earnings as reported by Statistics NZ's annual Income Survey, which from 2016 is undertaken as part of the Household Labour Force Survey.

Pay rates per hour are the best indicator of fair pay. CEVEP uses the average (mean) because this is the indicator that has been used to monitor progress since 1974, as well as for more technical reasons discussed below.

All statistical indicators have their limitations. In monitoring the gender pay gaps, CEVEP recommends focusing on the trend over time (see the graph on the Gender Pay Gap page). Progress has been very slow.

Statistical sources

Statistics New Zealand reports annually on the gender pay gap. It has done so since 1974, the year after the Equal Pay Act 1972 came into force.

Until 1996 data for this came from the Quarterly Employment Survey of employers. Employers report 'job level' data, including their average hourly pay for women and for men. See more on the QES below...

Since 1997, Statistics NZ's Income Survey, undertaken each June as part of the Household Labourforce Survey, has collected data on income by gender, ethnicity and age, and the distribution of income by quintiles. Each October, Statistics NZ published 'highlights' from the Income Survey, including gender pay gaps. The data tables from each survey, with average and median hourly and weekly earnings (Tables 10 and 11), can be downloaded from its website.

In 2016 this income survey was incorporated into the Household Labour Force Survey (with some slight technical changes), as was the Quarterly Employment Survey. Now known as Labour Market Statistics (income), data for each year from 1998 can now be found using the Statistics NZ's data hub table viewer.

This data comes directly from individuals aged 15+ in 15,000 households. The Survey asks employed people about their hours and pay rate in each job they have held in the past 12 months. It also collects data on income from other sources (self-employment, benefits, interest, dividends), as well as household composition and demographics. Taken together these can provide a picture of how New Zealanders are doing.

Statistics NZ says about the Income Survey:

Due to its large sample size, the NZIS is particularly good for making income comparisons across different population groups. For example, using the NZIS it is possible to compare incomes across demographic variables such as age, ethnicity, sex, qualifications and household type.

Quarterly Employment Survey

The QES was used to monitor the gender pay gap before the Income Survey, which is now Statistics NZ's best source for income data. Statistics NZ's website discusses the methods, merits and uses of its various surveys.

Some commentators continue to use the QES for pay gap data, and the media sometimes react to small quarterly changes in the QES. The QES is important for monitoring employment status and changes by industry. However, there are limits to the QES for gender pay gap purposes because:

The annual Income Survey resolves these issues. It asks employees themselves a much wider range of employment and demographic questions.

Another limitation is that QES occupational breakdowns are for broad sector categories only. In an increasingly casualised labour market, this hourly and weekly income data should not be used to calculate annual incomes - use Household Economic Survey and LEED person-level data for that.

Means and medians

Average hourly ordinary time earnings has been the indicator used to report on the gender pay gap since 1974. Despite the name 'pay gap', it is typically presented as the ratio of women's to men's average hourly earnings - for example, 86.1% in June 2014 (rather than 13.9%).

The Income Survey reports on both average and median hourly and weekly pay. The QES gives only the average. The mean or average is the sum of all pay divided by the number of earners. The median is what the middle person in pay range earns.

In June 2015 Statistics NZ's publication Measuring the Gender Pay Gap announced that it is adopting the median hourly rates as its measure, arguing that it is more 'typical' and less influenced by very high and very low earnings than the average (mean). Few very high earnings are captured by the Income Survey sample, so findings for this groups may be less accurate, or cause fluctuating findings year to year, says Statistics NZ.

CEVEP strongly disagrees with this change of indicator and has written to the Chief Statistician. New Zealand should stick with measuring the gender pay gap in average hourly ordinary time earnings for two reasons.

Firstly, because the change disrupts our monitoring of trends in the gender pay gap since 1974, and in gender/ethnicity gaps since 1997. CEVEP emphasis monitoring trends as small year-to-year changes in the percentage gap may be affected by shifts in (gendered) employment in different sectors, or by technicalities such as sample sizes in he early years of the Income Survey, and the immanent integration of the Income Survey into the Household Labourforce Survey.

Secondly, because very high earners and very low earners are an important part of the gender pay gap problem. Very high earnings are disproportionately white males - the CEOs, senior managers, professionals and academics in the NZ Census of Women's Participation, for example. Those on very low pay are disproportionately brown females - no shortage of them in the labour market. Monitoring the median woman as more 'typical' than the mean produces a narrower gap and smoother trend by dropping part of the problem - about 4 percent of it.

Following the market-oriented policy changes of the 1980s, the share of total income (from all sources) received by the top fifth and top tenth of earners began to increase disproportionately, while most wages fell against inflation and a reduction in resources for the poorest households in particular (Podder and Chatterjee 1998; Stillman et al. 2013).

Using average hourly data as the indicator for pay equity by sex and ethnicity means that narrow plume of high earners is taken into consideration, not excluded. That and the continuity of data back to 1974 are why average hourly ordinary time earnings are CEVEP's preferred indicator for monitoring the gender pay gap.

Comparing the indicators

Statistics New Zealand's website notes that the difference between the mean and the median reflects how wide income differences are. That is shown by the following graph which monitors the gender pay gap in both average hourly pay rates and median hourly pay rates over the period of the Income Survey. The difference between the two is growing, reflecting growing income disparity.

We add in the gender pay gap in average hourly earnings from the QES since 1974, and it becomes very clear why recently government ministers prefer to cite the gender pay gap in median hourly pay rates. It looks so much better. In June 2013 the median wage for women was 89.8% of the median for men compared to 87.3 for average hourly earnings - although in dollar terms the median wage nearly $4 lower than the average. In June 2014 the ratio of women's to men's median wage rate was 90.1%, compared to 86.06% comparing averages. In June 2015, the median gender pay gap still has a better look, though it dropped to 88.2% (for no very clear reason).

When the median rate for women (or men) is higher than the average, it means that half of all New Zealand women (or men) are earning less than the average wage rate per hour. The diverging average and median figures for the pay gap between women and men, since about 1998, suggest that for women in higher paid jobs and occupations the ratio has been improving, while women in many typical women's jobs continue to be undervalued relative to men. Caregiver Kristine Bartlett is an example.

In June 2015 Statistics NZ released a new report stating its preference for using the median, not the average, to measure the gender pay gap. CEVEP disagrees and corresponded with the department about this.

graph

Focus on trends

To monitor progress on pay equity, CEVEP recommends focusing on the trend shown by average hourly ordinary time earnings from the Income Survey. Small fluctuations in gender pay gap data, whether mean or median and from either source, may be the effect of changes in male employment or in industries, rather than improvement in female wages. It is important to consider the full picture shown by the latest statistics, and to focus on the trends in women's and men's pay over time.

In the graph above, the trend shows improvement when governments take action on pay equity.