PDF Data Dictionary

Data Collection

Labour Market Statistics: June 2020 quarter

Name
Labour Market Statistics: June 2020 quarter
Label
Labour Market Statistics: June 2020 quarter
Description

#Period-specific information

##Response Rates

 Survey

Reference period 

 Response rate

Sample rate

 HLFS

Each week during the quarter (1 April 2020 – 30 June 2019)

Target: 90 percent
Achieved: 74.3 percent

Target: 76 percent
Achieved: 69.8 percent

 QES

The pay week ending on, or before, 20 May 2020

Target: 89 percent
Achieved: 83.8 percent

N/A

 LCI

Pay rates at 15 May 2020

Target: 94 percent
Achieved: 95.2 percent

N/A

See New quality measures for the Household Labour Force Survey for more information on the sample rate and response rates.

##HLFS

###Impact of Covid-19 on collection period and response rates

Interviewing for the June-2020 quarter of the HLFS began on April 14 2020 while New Zealand was in alert level 4. Because of this, no face-to-face interviewing was allowed. First-time respondents in the HLFS are typically visited by a field interviewer, but this quarter they were instead sent a letter asking them to call the Stats NZ contact centre to complete their interview over the phone. This was the first time Stats NZ has used letters prompting respondents to call, so the effectiveness was unknown but it was anticipated letters would not be as successful as a visit from an interviewer.

In addition to field interviewers visiting first-time respondents, they also visit households that have been in the HLFS previously but the contact centre has been unable to reach in the current quarter. There were concerns that the response rate and achieved sample rate would suffer this quarter because field interviewers were unable to make house visits.

Interview times for the HLFS were significantly longer this quarter. The income supplement and disability supplement were included, as they are every June quarter, and there was also a well-being supplement and extra questions (surrounding working from home and job security) added to the HLFS. Longer interview times have negative effects on response rates.

The decision was made to re-contact respondents who should have otherwise had their final HLFS interview in the March-2020 quarter, as an effort to improve the response rate and the achieved sample rate.

Field interviewers were able to resume house visits for the last two weeks of collection, beginning on Monday 22 June. The final day for interviewing was Sunday 5 July.

###Response rate

With the inclusion of respondents doing a ninth HLFS interview, the overall achieved sample rate was 69.8 percent. This is below the target of 76 percent, and is the lowest ASR since the September-2012 quarter. However, the ASR of 69.8 percent is not much lower than achieved sample rates seen in previous recent June quarters (70.3 percent and 70.9 percent in June-2017 and June-2018, respectively), and only slightly lower than last quarter’s 70.7 percent.

###Additive outlier treatment

X13 is the tool used by Stats NZ to perform seasonal adjustment. If a value is sufficiently extreme due to a short-term break in the raw data, X13 may not fully account for the outlier. Tests and analysis significantly indicate that the series listed below were impacted by the unusual events experienced this quarter related to Covid-19 and the lockdown:

  • Male, female and total employed
  • Male, female and total unemployed
  • Male, female and total not in the labour force (NILF)
  • Employed, aged 15-64 years
  • Unemployed, aged 15-64 years
  • Not in the labour force, aged 15-64 years
  • Total actual hours worked

The June-2020 quarter has been treated as an additive outlier for these series. The outlier treatment effectively removes the June quarter from seasonal factor calculations and prevents the June quarter from causing undesirable revisions to current and historical seasonal factors.

Trend values for the June-2020 quarter are not being published, as the calculated trend value is not accurate this quarter. Trend series are slow to react to change, and will behave as if there were no Covid-19 or lockdown for some time. Reporting trend estimates will likely underestimate the effect of the pandemic.

##QES

###Response rates

Low response rates were anticipated in the QES for the June 2020 quarter, due to the impacts of Covid-19 and lock down. There was the potential for heavy business churn, respondent unwillingness to cooperate in times of economic stress, and the concern that hardest hit businesses would be least likely to respond, thus biasing the responses we did receive. In addition, the QES is primarily paper based, which could have been particularly vulnerable to level 4 lockdown restrictions on the post for non-essential businesses.

We were able to mitigate this by:

  • providing an electronic form to all businesses with valid email addresses on record in addition to the regular paper form;
  • introducing courtesy pre-reminder calls;
  • targeting responses by broad industry and sector group.

Additionally, the swift end to lockdown and resumption of normal business gave respondents time to return to the office before responses were required on July 10th, 2020. As a result, the final response rate was 83.8 percent, only 4.6 percentage points below the same quarter in 2019. Upon further assessment, we concluded that this was acceptable and no changes to our standard imputation were required.

###Response quality

Another early concern in the QES was that the responses received would be poor quality, or full of partial completes. This was mitigated by the inclusion of additional information on the cover letter and FAQs for respondents, and guidance and extra support provided for processors and call centre staff.

This advice was as follows: “How do I record changes due to Covid-19? If your business is receiving a wage subsidy, please report the amount paid including the wage subsidy when completing the survey.

If you are still paying employees who are unable to work, please record hours paid at their contracted hourly rate.

If you are paying reduced wages, but you have not negotiated a reduced hourly rate, an employee’s hours paid should be reduced. An employee who usually earns $500 per week for 40 hours work but is now earning $400 per week would have reduced hours paid of 32. If you have agreed a reduced hourly rate please record hours paid as you normally would.

Please use the comments box to report changes due to Covid-19 as you would to report other changes. These changes may include:

  • Voluntary pay decreases. Please provide details if any staff have volunteered to take decreased pay or reduced hours.

  • Temporary or permanent redundancies. Please record how many positions have been affected and for how long.

  • Your business being temporarily or permanently closed. Please provide details on how long your business has been closed and if you intend to reopen.”

While there have been some quarter specific response challenges, the majority of respondents seem to have understood the guidance given, and provided appropriate data. We are confident that our final estimates are of good quality and accurately reflect the state of New Zealand’s labour market in the June 2020 quarter.

###Interquarterly changes in sample errors

There were a large number of extreme changes in absolute sample errors over the quarter. This is an understandable reflection of extreme real-world quarterly changes but might have negative impacts on interquarterly comparisons. As a result, we recommend – as ever – focussing on annual changes in unadjusted estimates.

###Seasonal adjustment and trend series

Covid-19 caused many abrupt changes to QES time series. This was a challenge for the creation of seasonally adjusted and trend series and caused us to consider the possibility of explicitly marking this quarter as an outlier. After further analysis, however, we could not conclude that this would be more appropriate than our standard automatic outlier process, and so decided not to manually adjust either series. This decision will continue to be reviewed in subsequent quarters.

##LCI

###Response rates

The target response rate for the labour cost index is 94 percent (100 percent for key firms, 94 percent for non-key firms). The labour cost survey is primarily a postal, paper-based survey. Surveys for the June 2020 were sent out with reference date of May 15th 2020. The end of lockdown and move to alert level 2 meant that respondent businesses were able to report their employees’ pay at the reference date. As a result, response rate was 95.2% (above the target response rate of 94%).

###Response quality

In March 2020 the Government announced the twelve-week Covid-19 Wage Subsidy Scheme, to support businesses affected by Covid-19 restrictions. We anticipated that businesses that received the subsidy may find it confusing to report employees’ wages for the LCI.

To mitigate this, we:

  • included additional information along with the LCI survey to guide and support;
  • increased telephone calls with respondents to follow-up on and confirm wage changes, as well as the reasons behind them.

For electronic survey respondents and the FAQ sheet attached to postal surveys, we provided the guidance below:

How do I record changes in labour costs due to Covid–19?

If your business is receiving a wage subsidy, please report the amount paid to each employee including the wage subsidy when completing the survey. Please use the Reasons for change comments box (Column E) to report changes due to Covid-19 as you would to report other changes in labour costs. These changes may include:

  • Receiving a wage subsidy. Please record details of which positions are receiving a subsidy and the amount paid to each position excluding the subsidy. Use the main survey to record the amount paid to each employee including the subsidy.
  • Temporary or permanent redundancies. Please record details of which positions have been affected and for how long.
  • Your business being temporarily or permanently closed. Please provide details on how long your business has been closed and if you intend to reopen.