Introduction
The COVID-19 pandemic in Ireland has had a significant impact on the collection of data for the Earnings and Labour Costs (ELC) release. Preliminary estimates for the first Quarter (Q1) of 2020 and Final estimates for Quarter 4 (Q4) 2019 are being published tomorrow, 9 June 2020. The aim of this technical note is to outline to ELC users the impacts from COVID-19 on the collection of data for the ELC release and to outline what impact they can expect to see for the duration of the COVID-19 crisis.
Earnings and Labour Costs (ELC) Background Information
The ELC release is published based on data collected by the Earnings, Hours and Employment Costs Survey (EHECS). The EHECS collects data from enterprises with three or more employees in the NACE Rev 2 Industrial Sectors B–S inclusive.
The following are the groupings of NACE industrial sectors that are used when presenting ELC data:
Sector Group Sector Name
B-E Industry
F Construction
G Wholesale & retail trade: repair of motor vehicles & motorcycles
H Transportation & storage
I Accommodation & food services activities
J Information & communication
K-L Financial, insurance & real estate activities
M Professional, scientific & technical activities
N Administrative & support services activities
O Public administration & defence
P Education
Q Human health & social work activities
R-S Arts, entertainment, recreation & other service activities
All enterprises with 50 or more employees and a sample of those with 3 to 49 employees are surveyed as part of the EHECS each quarter.
The information collected from an enterprise as part of the survey includes:
Imputation is carried out for non-responding enterprises with 50 or more employees. Where an enterprise responded in at least one of the previous four quarters ratio-imputation is used to estimate figures for the current quarter. For enterprises that did not respond in any of the previous four quarters a stratum average imputation method is used to estimate the missing variables. These estimates are based on respondent enterprises of a similar size and activity.
For enterprises with 3 to 49 employees inclusive, a weighting factor (the reciprocal of the sampling fraction) is used to weight the estimates up to the total population for both employees and enterprises. Enterprises with 50 or more employees are assigned a weight of 1 as the sample consists of a census of these enterprises.
All enterprises' data are contained in the quarterly dataset which is tabulated to produce the aggregates which are published in the ELC quarterly release. Results for the most recent quarter are published initially as “Preliminary” estimates and are revised and relabelled as “Final” when the next quarterly release is published three months later.
COVID-19 Income Supports
During March 2020, two major schemes were introduced by the Government to support those whose income from employment had been affected due to COVID-19.
The majority of those whose income from employment was affected due to COVID-19 in Q1 2020 were being facilitated through the COVID-19 Pandemic Unemployment Payment which is administered by the Department of Employment Affairs and Social Protection (DEASP).
A smaller cohort were being facilitated through the Revenue Temporary Wage Subsidy Scheme (TWSS). The scheme enables employees, whose employers are affected by the pandemic, to receive significant supports directly from their employer through the payroll system. This was initially introduced as the Employer Refund Scheme (ERS) on 15 March and enabled workers who are temporarily laid off due to the COVID-19 pandemic to claim a special support payment of €203 per week. The ERS was replaced by the TWSS on 24 March. The scheme was expected to last 12 weeks from 26 March 2020 but has now been extended until the end of August 2020 (See Revenue website for more information on the TWSS scheme).
The Central Statistics Office (CSO) have engaged extensively with both Revenue and the DEASP to understand the mechanics behind the various income supports. The CSO have also engaged with Eurostat, the European statistical agency, to determine the statistical treatment of these income supports for the purposes of the collection of data for EHECS and the reporting of the related outputs.
Any payments related to the Pandemic Unemployment Payment scheme are not collected by EHECS or recorded in the ELC release. Such payments are paid by DEASP to a person who has lost income from employment and where no attachment in the form of an employment contract is maintained between the employer and employee. Eurostat has confirmed that such payments should not be recorded in labour costs statistics.
TWSS payments are recorded as follows in the ELC release:
Effect of COVID-19 on data collection and response rates from the Earnings, Hours and Employment Costs Survey (EHECS)
Data for Earnings, Hours and Employment Costs Survey (EHECS) is collected quarterly at enterprise level. Questionnaires were issued to enterprises at the beginning of April 2020.
Due to the unprecedented emergency, Government measures were put in place in response to COVID-19, which resulted in the sudden closures of non-essential businesses in weeks 11 – 13 of Q1 2020. These restrictions were ultimately extended and affected the ability of a large share of enterprises to respond during the EHECS data collection period.
The CSO extended the data collection period for Preliminary estimates for Q1 2020 for an additional two weeks to maximise response rates. The overall response rate for the Preliminary estimates for ELC Q1 2020 is 43%. This contrasts with the comparable response rate for Preliminary estimates of 53% for ELC for Q4 2019 and of 54% for ELC for Q1 2019. Because of the low response rates, 48% of enterprises required imputation for Q1 2020 and caution is therefore advised in relation to drawing inferences from the Preliminary estimates for Q1 2020.
We understand the difficulties enterprises faced in both accessing and sending us the required data and acknowledge the efforts made by our survey respondents during this time. The CSO is hopeful that businesses who have been unable to submit their Q1 2020 return yet will be able to provide this data in the coming weeks for inclusion in the Q1 2020 Final ELC Release, scheduled for publication in August 2020.
COVID-19 Impact on methodology for the ELC release
The Earnings Analysis team within the CSO have engaged with Eurostat and colleagues from the Methodology division within the CSO to determine what changes were required to the usual methodology that is required for the ELC release due to the impacts of COVID-19.
The ELC release presents aggregate results which are based on enterprises’ survey returns. Due to the introduction of the TWSS at the end of March 2020, many enterprises were unable to record the refunds received in relation to the scheme. As a result, EHECS data was linked with an administrative data source from the Revenue Commissioners that contains information about the TWSS scheme. The linkage and analysis were undertaken by the CSO for statistical purposes in line with the Statistics Act, 1993 and the CSO Data Matching Protocol.
Definitions of the variables included in the ELC release are available in the background notes for the release which are also available separately here.
Registered employment refers to the employment of respondent enterprises at the end of the reference quarter.
Average employment refers to the average of the quarters’ opening employment and closing employment. This variable is used as the denominator in calculating average earnings per person and average paid hours per person.
Average weekly earnings are calculated by dividing the total earnings for the quarter by the average number of persons employed during the quarter and then dividing that by 13 (the number of weeks in a quarter).
Government measures put in place in response to COVID-19 related to weeks 11 – 13 of Q1 2020 for a limited number of sectors. As a result, total wages and salaries and hours worked by all employees in an enterprise related to most, if not the entirety, of the quarter. For enterprises in certain sectors the closing employment at the end of the quarter was significantly reduced and resulted in the skewing upwards of average hours and average earnings in that sector. To overcome this issue, a weighted average employment figure amounting to 11/13 of the employment at the start of the quarter and 2/13 of the employment at the end of the quarter was used as a denominator in calculating average earnings and average paid hours. This weighted average employment denominator was only used for NACE sectors 56 (Food and beverage service activities) and 96 (Other personal service activities) for the Preliminary estimates for Q1 2020.
Example: Enterprise ABC had 10 employees at the start of Q1 2020. These 10 employees worked 40 hours weekly for 11 weeks of Q1 2020 until Government restrictions meant the enterprise had to close. All 10 employees were laid off and sought income support in the form of the Pandemic Unemployment Payment from DEASP. The total number of paid hours for the quarter is 4,400 (10 employees worked 40 hours per week for 11 weeks each: 10 x 40 x 11 = 4,400). The average weekly paid hours is calculated by dividing the total paid hours for the quarter by the average number of persons employed during the quarter and then dividing that by 13 (the number of weeks in a quarter). Using an average employment of 5 (where there were 10 employees at the start of the quarter and zero remaining at the end of the quarter), average weekly paid hours equals (4,400/5)/13, which amounts to average weekly paid hours of just under 68 hours. If we consider that the 10 employees were likely to have worked for 11 weeks of the quarter we get a weighted employment of 8.46 employees over the quarter ([(10x11)+(0x2)]/13). If we use the weighted average employment of 8.5 in the calculation of the average weekly paid hours, i.e. (4,400/8.46)/13, the average weekly paid hours equals 40 hours. Thus, even though there might be less hours actually worked in Q1 2020, due to COVID-19, the average weekly paid hours will not capture the reduction in hours due to employees being laid off and being put on the Pandemic Unemployment Payment scheme because the denominator is also scaled down.
Similar to the example above, if the 10 employees from ABC were placed on the TWSS scheme instead of the Pandemic Unemployment Payment scheme the information returned under EHECS will not indicate whether they continue to work or not but rather just that they are continuing to be paid and have an attachment to their employer.
Because of the statistical treatment we are obliged to follow for the COVID-19 income support schemes the results presented in the ELC release will not capture fully reduced hours worked due to COVID-19. We will separately look to carry out additional analyses of hours data and present any results independently of the ELC release.
Q1 Final and Q2 Preliminary 2020 results – due for publication August 2020
The Earnings and Labour Costs Q1 2020 (Final) and Q2 2020 (Preliminary Estimates) release is currently scheduled to be published on 26 August 2020 and, as usual, any amendments to the schedule will be signalled on the CSO release calendar.
Government measures put in place in response to COVID-19 related to weeks 11 – 13 of Q1 2020. The impact of the pandemic on earnings and labour costs results for Q1 2020 is minimal as a result, as most industrial sectors were unaffected for 11 weeks of the quarter, with some sectors unaffected for the entirety of the quarter. It is anticipated that a more significant impact will be evident in Preliminary estimates for Q2 2020 due to the continued closure of a large number of businesses throughout the full survey period, i.e. April to June 2020.
Previous reviews of Preliminary and Final estimates indicated that Preliminary figures can be accepted as a suitable indicator for quarterly trends. Analysis of the results for individual NACE industrial sectors highlighted that the change from Preliminary to Final estimates was broadly in the range of plus or minus 5%. However, response rates can significantly impact on the quality of data at the sectoral level. It is therefore possible that, because of the lower response rates seen for the Preliminary estimates for Q1 2020, there will be a more significant change between the results being published as Preliminary tomorrow and those published as Final later this year.
We understand the difficulties that enterprises have faced and wish to thank all survey respondents who supplied their data and engaged with the CSO in relation to the difficulties they faced. We appreciate the continuing efforts of survey respondents to get information to us and it is by collecting survey information that we will be able to report on the effects of COVID-19 on our economy.
If any users have any questions or need any clarification in relation to anything outlined in this Technical Note or the Earnings and Labour Costs release please contact us at earnings@cso.ie
For information, this technical note was issued on 08 June 2020.