Italian National Institute of Statistics (ISTAT)

Integrated Labour, Education and Training Division

Social Statistics and Welfare Directorate

Statistics Production Department

Istat - Italian National Institute of Statistics

Via Cesare Balbo, 16 - 00184 Rome - Italy

Indices of volumes of hours worked by employees (STSIND, STSCONS, STSRTD, STSSERV).

NACE Rev. 2

Activities covered:

Annex A: divisions 06-36 and sections B,C,D;

Annex B: section F;

Annex C: sections G and divisions 45-46-47 (and G47XG473) ;

Annex D: divisions 49-53 and divisions 55-56 and divisions 58-63, H, I, J, L, M_STS, N.

Size classes covered:      
Enterprises with at least one employee.

Hours worked by employees are defined in coherence with the EBS Regulation (REGULATION (EU) 2019/2152 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 27 November 2019 on European business statistics). In particular, on the basis of COMMISSION IMPLEMENTING REGULATION (EU) 2020/1197 of 30 July 2020 laying down technical specifications and arrangements pursuant to Regulation (EU) 2019/2152, total hours worked by employees:

represents the number of hours actually worked by employees, for the output of the statistical unit during the reference period. Time spent on adjacent work, indirectly contributing to the output (e.g. planning, preparation, administrative and alike), as well as time spent without actual work, but deemed and remunerated as such by the statistical unit (e.g. short breaks, short disruptions due to slack in production, trainings and alike) is included. L 271/92 Official Journal of the European Union 18.8.2020 EN Time spent on work, be it adjacent work, without actual remuneration (e.g. unpaid overtime) is also included. Remunerated time spent without actual work and not deemed as such by the statistical unit (e.g. annual leave, sick leave, maternity leave, official holidays, longer breaks, meal breaks, strikes, commuting and alike) is excluded. Infra-annual statistics may not be able to take into account all these items such as unpaid overtime

Reporting and observation unit is enterprise

All enterprises with at least one employee which were active in the reference quarter in the STS economic activities. In 2022, these enterprises were, on average over the four quarters, 1.217 million.

The whole country. Activities outside the geographical coverage are not included in the data.

The series start in 2000Q1.

2015

Data both trasmitted to Eurostat and published on a national level are Indices with base year 2015=100.

Year 2022. Quarterly data.

Legal basis: In addition to the EBS Regulation (REGULATION (EU) 2019/2152 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 27 November 2019 on European business statistics), all the three sources involved in the production of the indicators are among the compulsory surveys identified every year by a Decree of the President of the Republic (DPR).

Obligation on units to provide data:

The two business surveys are mandatory. Penalties are foreseen for non-response of units with at least 100 employees.

Furthermore, firms have an obligation to fill monthly the social contributions forms on which the administrative source is based.

Planned changes in legal basis, obligation to respond and frame used: None.

None. (Data are not sent to OECD, UN, etc. or used in reports sent to these institutions). 

Links to legal acts on the production of official statistics and confidentiality can be found on the Sistan - National Statistical System website (the Sistan website is only available in Italian)

Sections and divisions of all Annexes are free

At the moment the data are not released.

At the moment the data are not released.

Data are transmitted to Eurostat via Edamis.

Both trasmission to Eurostat and national dissemination take place on a quarterly basis.

At the moment the data are not released.

Not available

At the moment the data are not released.

At the moment the data are not released.

Data are trasmitted to Eurostat quarterly, within the 90 day regulation deadline, in SDMX format.

As indicated below in 18.1, volumes of hours worked by employess in 1+ employee firms are calculated multiplying per capita hours worked by employees on 10+ employee firms by the number of jobs in 1+ employee firms.

There are no methodological publication describing the method employed to calculate volumes of hours worked by employees in 1+ employee firms.

However, the methods used to calculate per capita hours worked by employees on 10+ employee firms are described in the following publication: Bellisai, D., Binci, S., Gigante, S., Libratore, A., Serbassi, L. and M. Sorrentino, ¿Verso la modernizzazione della produzione delle statistiche congiunturali sull¿input di lavoro: il processo di integrazione tra la rilevazione mensile sulle grandi imprese e quella trimestrale sui posti vacanti e le ore lavorate¿, Istat, Working Papers, n.11, 2013 (http://www.istat.it/it/files/2013/12/IWP_11_2013.pdf). 

Not available.

The data are produced in accordance with the European Statistics Code of Practice and with its Italian version (http://www.istat.it/it/files/2011/11/codice_statistica.pdf).

A specific assessment of data quality for each quality dimension is presented below. There are no relevant quality problems that deserve mentioning here.  

Eurostat is the main user.

The produced data are coherent with the requests of the EBS Regulations (REGULATION (EU) 2019/2152 and COMMISSION IMPLEMENTING REGULATION (EU) 2020/1197). 

The data are considered satisfying the requests by Eurostat included in the EBS Regulations (REGULATION (EU) 2019/2152 and the COMMISSION IMPLEMENTING REGULATION (EU) 2020/1197).

All STS requirements are fulfilled. 

As the HOUES series are calculated by multiplying survey based per capita hours worked by administrative based jobs, sampling errors affect HOUES via per capita hours worked. For an assessment of sampling errors on the numerator and dominator of per capita hours worked, see 13.2.

For what concerns non sampling errors, response rates can be found below in 13.3, while descriptions of models used in editing and imputation and grossing up can be found below in 18.5. Coverage errors affect the per capita hours worked by employees because the surveys¿ reference and study population differ due to the enterprises beginning and finishing their activities in the reference population between the SBR reference year and the surveys reference quarter. If these errors are measured via the difference between the SBR used by the survey and the OROS (see below, type of source, 18.5) list for the reference quarters, for 2021:

- 18.5% of firms in the survey's reference SBR were not in the OROS list for the reference quarters, and

- 23.9% of firms in the OROS list for the reference quarters were not in the surveys's reference SBR.

As the HOUES series are calculated by multiplying survey based per capita hours worked by administrative based jobs, sampling errors affect HOUES via per capita hours worked by employees.

For what concerns survey based per capita hours worked by employees, in the four quarters of 2022 the coefficient of variation is estimated as ranging between 0.3 and 5.0% at the section level in the Nace sections covered by the EBS Regulations (REGULATION (EU) 2019/2152 and COMMISSION IMPLEMENTING REGULATION (EU) 2020/1197).

Response rate for VELA and LES, considered together:

quarter                             respondents                 sample               response rate

2022Q1                                8,341                     11,939                   69,9

2022Q2                                8,329                     11,935                   69,8

2022Q3                                8,626                     11,936                   72,3

2022Q4                                8,354                     11,927                   70,0

2022 (average)                    8,413                     11,934                   70,5

Actions to speed up or increase the rate of response:

Monthly reminders (by e-mail) and intensive follow-ups by phone are addressed to non responding LES units. Once a year a warning with penalty (registered letter with return receipt) is sent to firms that have not answered to LES during the previous year.

Quarterly reminders by phone and e-mail are made to the VELA survey non responding units. Once a year a warning with penalty (registered letter with return receipt) is sent to firms that have not answered to VELA during the previous year and have at least 100 employees.

On coverage errors, see 13.1.

Timeliness: data for 2022 have been transmitted to Eurostat on average 82 days after the end of the reference quarter, that is 8 days before the regulation deadline.

Timetable of data collection:  for LES, the deadline for responses is fixed at 18 days after the reference month. For VELA the deadline for responses is fixed at 37 days after the reference quarter.

Data for 2022 have been transmitted to Eurostat on average 82 days after the end of the reference quarter, that is 8 days before the regulation deadline.

Hours worked by self-employed are not included. On all other aspects, hours worked by employees are defined in coherence with COMMISSION IMPLEMENTING REGULATION (EU) 2020/1197. The data cover only the entire national territory. 

The 2000-2003 data are based on data collected through LES only (and not also through VELA) and appropriate estimation procedures that make the per capita hours worked by employees that are multiplied by the OROS jobs (see below, type of source, 18.5) representative of all enterprises with 10+ employees.

HOUES data are coherent with those transmitted to Eurostat for STS variables EMPL and WAGE, because the series for all three variables are based on the OROS population of firms active in the reference quarter and measure of jobs.

Data at different aggregation levels are coherent.

For 2022:

-  for VELA, the cost of the phone reminders was equal to 183,000.

Furthermore, it can be estimated that around 5 person hours are needed each quarter to produce the HOUES indicators (from data collection, to data processing and validation).

Under the assumption that around one hour of work is needed to fill in the relevant parts of administrative forms and survey questionnaires by each responding firms each quarter, the burden on respondents can be estimated at around 1.217 million of hours of work by  employees per quarter. 

The data are disseminated provisionally at the time of the first transmission to Eurostat. A revision of the data will be carried out yearly with the dissemination of the indices of the first quarter for the period covering the previous 8 quarters.

If errors are identified during the year:

-          if they are considered relevant, the series are re-calculated and resent as soon as possible;

-          if they are considered less relevant, the series are re-calculated and resent together with the following regular revision, that is with the data for the first quarter of the year.

So far, comparisons with other sources has not led to revisions.

The revisions carried out once a year are due to late respondents in the survey and administrative sources.

A vintage database of all releases is available.
Istat's principles, rules and classifications of short-term statistics revisions can be found here: http://www.istat.it/en/economic-trends/revisions.

Type of source:

volumes of hours worked by employees in 1+ employee firms are calculated multiplying per capita hours worked by employees on 10+ employee firms by the number of jobs in 1+ employee firms. Per capita hours worked by employees are based on two business surveys: VELA, a quarterly survey on 10-499 employee firms; and LES, a monthly survey on 500+ employee firms. The number of jobs in 1+ employee firms is derived from an administrative source (OROS) and is the same transmitted for the STS EMPL variable (detailed information on this source can be found in the metadata on the EMPL variable).

Frame on which the source is based:

for both VELA and LES the latest available release of the SBR.

Sample or census:

VELA is sample based; LES is a census.

Criteria for stratification:

for LES none.

Economic activity, size and geographical area are used as stratification variables for the VELA sample. The considered stratification classes are the following:

- for economic activity: divisions from 05 to 36 and from 45 to 47; the aggregate of divisions 37, 38 and 39; sections F, and from H to N;

- for size (in terms of employees): 10-19, 20-99 and 100-499;

- for geographical area: North, Centre and South (this last class includes Sardinia and Sicily).

Threshold values and percentages:           

In 2022, the sample sizes for the STS domains were the following:

- for LES, around 1,200;

- for VELA, around 11,000, representing 7.0% of the population of 10-499 employee firms in STS domains in the latest release of the statistical business register

- for OROS, the population size is around 1.217 million.

Frequency of updating the sample:

Once a year (for the wave referring to the first quarter):

- the list of all the population firms with at least 500 employees is updated on the basis of the SBR most recently available version, and

- around one third of the 10-499 employee sample firms is rotated. 

Frequency for per capita hours worked by employees (business surveys): for 10-499 employee firms, quarterly.

Frequency for 500+ employee firms, monthly.

Frequency for jobs (administrative source): monthly.

Questionnaires used in the survey:

See attached files:

BCS_HOUES_A_IT_2022_0000_an1 (VELA survey) and BCS_HOUES_A_IT_2022_0000_an2 (LES survey).

 Data collection media:  web, administrative data.

It is carried out through interactive assessment of influential observations, which are identified via both automated procedures and experts¿ analyses on aggregate data. 

The files that are sent to Eurostat are produced from data stored in an Oracle database via a generalised Istat software. After their production, they are not checked with any further software or specialized tool. 

Estimates for non-response:

for what concerns LES, the methodology of estimating missing data (unit and item non responses) works on a deterministic basis. It uses both information of clusters defined in terms of economic activity and time series data of the enterprise itself.

For what concerns the VELA survey, only item non responses are imputed. For what concerns jobs (at the denominator of per capita hours worked), editing and imputation is based on an auxiliary source, the OROS survey (based on the social security database of the forms that all enterprises with at least one hour of remunerated work have to fill each month to pay the compulsory social contributions). For what concerns hours worked, editing and imputation of item non responses includes: identification of outliers and selective treatment of the most relevant ones; deterministic editing procedures; donor imputation procedures based wherever suitable on the same quarter of the previous year data for the same firm on which they are applied; identification and analysis of the records most influential for the aggregate data.

Estimates for grossing-up to population levels:

all responding firms and the unit non responses imputed records are assigned a weight via calibration. The only exception is a small list of extremely large firms which are assigned a unit weight.

The theoretical reference population for the calibration is made by all the enterprises with at least 10 employees in Nace Rev. 2 sections B to N which were active in the reference quarter. This population is well represented by the set of microdata of the OROS survey for the reference quarter. It is this set of microdata that is therefore used to derive the calibration constraints. The quarterly average of monthly data on jobs, as measured by OROS, is used as the auxiliary variable in the calibration.

The calibration classes are based on economic activity and enterprise size. The economic activity classes for calibration are based on Nace Rev. 2 divisions for sections from C to E, G and N, while they are based on sections for B, F, and from H to M. The considered size classes are based on employees and are thus defined: 10-19, 20-99, 100-499 and at least 500.

The initial calibration weights are based on both inclusion probabilities and response rates.

The calibration is carried out using a generalized software purposely built by Istat, ReGenesees, and within it, a truncated logarithmic distance function.

Type of index: value indices with a fixed base.

Method of weighting and chaining: none.

Planned changes in production methods: none.

See attached file: BCS_HOUES_A_IT_2022_0000_an3.

Not available