Reference metadata describe statistical concepts and methodologies used for the collection and generation of data. They provide information on data quality and, since they are strongly content-oriented, assist users in interpreting the data. Reference metadata, unlike structural metadata, can be decoupled from the data.
The quarterly data are broken down by economic activity (at section level) in accordance with NACE Rev. 2 - Statistical classification of economic activities in the European Community.
2.3. Coverage - sector
All economic activities defined by NACE Rev. 2, except the activities of households as employers and the activities of extraterritorial organisations and bodies.
2.4. Statistical concepts and definitions
A 'job vacancy' is defined as a paid post that is newly created, unoccupied, or about to become vacant:
for which the employer is taking active steps and is prepared to take further steps to find a suitable candidate from outside the enterprise concerned; and
which the employer intends to fill either immediately or within a specific period of time.
‘Active steps to find a suitable candidate’ include:
notifying the job vacancy to the public employment services,
contacting a private employment agency/head hunters,
advertising the vacancy in the media (for example internet, newspapers, magazines),
advertising the vacancy on a public notice board,
approaching, interviewing or selecting possible candidates/potential recruits directly,
approaching employees and/or personal contacts,
using internships.
'Specific period of time’ refers to the maximum time the vacancy is open and intended to be filled. That period shall be unlimited; all vacancies for which active steps are continuing on the reference date shall be reported.
An 'occupied post’ means a paid post within the organisation to which an employee has been assigned.
2.5. Statistical unit
Local unit.
2.6. Statistical population
All establishments with at least one or more employees.
2.7. Reference area
All establishments located in the territory of Germany.
2.8. Coverage - Time
Quarterly data from 2010Q4 onwards.
2.9. Base period
Not applicable.
3.1. Source data
Identification of the source of the data
IAB-Job Vacancy Survey
Coverage
Geographical
The whole territory of the country
NACE
NACE sections A-S
Enterprise size
Establishments with one and more employees
Definition of the statistical unit
Local unit
Sampling design
Base used for the sample
For the postal survey in the fourth quarter of 2024, a sample of about 140,000 local units is drawn randomly from the official business register of the Federal Employment Agency. The sampling ratio is 6.6 % of the population of local units with registered employees in Germany (see Table 1).
The sample is stratified by:
2 regions (East and West Germany)
24 sectors (NACE Rev. 2 since IV/2010)
7 size classes
The number of local units in every cell of the sample matrix is chosen regarding the participation rates in previous surveys to ensure a sufficient number of responses. The probability of having an open vacancy is especially low in very small units. On the other hand, very small units account for about 30 per cent of all vacancies. For this reason, the absolute sample size is larger in smaller-sized classes (see Table 1).
The net sample size of the postal survey in the fourth quarter of 2025 was 15,905, thereof 11,311 in the sample for West Germany and 4,594 in the sample for East Germany (Table 2).
From the 4th quarter of 2023, the invitation for participating organisations was changed and now focuses more on online participation. After the initial invitation to participate online, one reminder with the paper questionnaire enclosed was sent to the companies that had not yet responded. The new procedure contributed to a higher response rate compared to previous years; the share of online responses has also increased.
In addition to the 4th quarter, this new mode was also used for the first three quarters of the following year.
Due to a high response rate in the 4th quarter of 2023, the samples for the follow-up surveys in the first quarter of 2024 consisted of around 18,000 local units. These are subsamples of the previous year's main survey (Tables 3 to 5). The net sample size in the first three quarters of 2024 averaged over 12,000.
The four surveys thereby formed a quasi-panel from the 4th quarter to the 3rd quarter of the following year. In the observation period 2023/24, 44.3 % of participants took part in all four surveys, 16.8 % in at least three, and 13.9 % in at least two of the surveys (Table 6). 25 % only participated in the main survey.
The response rate for the main survey in 2024 was 11.4%, and for the follow-up surveys, it ranged from 69.7% in the first quarter to 66.1% in the third quarter of 2024.
Sampling design
The number of local units in every cell of the sample matrix is chosen regarding the participation rates in previous surveys to ensure a sufficient number of responses. The probability to have an open vacancy is especially low in very small units. On the other hand, very small units’ account for about 30 percent of all vacancies. For this reason, the sample size is larger in small size classes (Table 1).
Retention/renewal of sampling units
The sample is renewed once a year in the fourth quarter.
Sample size
Gross sample size:
Q1 2024: 18,268 local units
Q2 2024: 18,190 local units
Q3 2024: 17,865 local units
Q4 2024: 139,247 local units
Net sample size:
Q1 2024: 12,725 local units
Q2 2024: 12,527 local units
Q3 2024: 11,783 local units
Q4 2024: 15,905 local units
Stratification
The samples for Eastern and Western Germany are stratified by
24 sectors
7 size classes
Other sources
Maintenance agency
The survey is conducted under the responsibility of the Institute for Employment Research (IAB) at the German Federal Employment Agency. On behalf of the IAB, the 2024 survey was conducted by Economics & Data ED23 GmbH in Munich.
Quarterly average over all business days of the respective quarter.
3.3. Data collection
Brief description of the data collection method(s)
Remarks
There was a methodological shift in how participants were invited, beginning in the fourth quarter of 2023. Respondents were initially invited to take part in the online survey only in October 2023, while a paper copy was only sent with the reminder invitation in November 2023. Previously, the paper survey was issued twice: first with the initial invitation and then with the reminder. There was also an option to download the personalised blank questionnaire in PDF format if needed.
Unlike previous years, the methodology introduced in the fourth quarter of 2023 was also adopted for the follow-up surveys (of each wave). Starting in the first quarter of 2024, CATI-based telephone interviews were replaced by the online/paper methodology already used in the fourth quarter questionnaires. Prior to this switch, a mode experiment indicated that responses and answers were not affected by such a change.
All surveys use standardised questionnaires that contain the same core questions to measure job vacancies and use the same definitions.
The first mailing was sent in October. Respondents were initially only invited to participate in the online survey. The second mailing, which was only aimed at non-respondents, was sent in November and included a paper copy of the survey. However, the recipients were able to respond online at any time.
There was also the option to download the personalised blank questionnaire in PDF format, if the company needed it to collect the requested information or preferred to answer on paper.
In the fourth quarter of 2024, 83.2% of the respondents participated online, while only 16.8% responded on paper.
The follow-up surveys in the first three quarters of 2024 are only aimed at respondents from the main survey in the fourth quarter of 2023. In contrast to the earlier telephone interviews, the follow-up surveys also used the new methodology (paper/online).
3.4. Data validation
Not applicable.
3.5. Data compilation
Brief description of the weighting method
Weighting dimensions
In 2015, a new extrapolation procedure based on the Generalised Regression Estimator (GREG) was developed in close collaboration between the IAB, the Economix Research and Consulting survey research institute and the Regensburg University of Applied Sciences. This estimator is widely used in modern survey research (cf. Särndal et al. 1992; Deville/Särndal 1992) and is also recommended by the European Commission (cf. European Commission 2002).
The GREG procedure continues the basic ideas of the extrapolation method used in the past. However, the adjustment to registered vacancies as an anchoring variable has been discontinued. Now, the number of employees subject to social security contributions and the number of establishments are used as anchor variables. Note that we base these numbers on the most recent available data, which usually lags by 2-3 quarters. A forecast of employment development is used to adjust the anchor variable to the current employment levels in the German economy.
The method allows the implementation of different kinds of weights. It allows considering non-response corrections and the direct analytical calculation of key statistics to assess the quality of the extrapolation and the validity of results, which allows statements regarding the survey's statistical validity. The GREG procedure is more efficient in methodology than the previously used procedure. It significantly improves the quality of the extrapolation.
GREG Implementation:
After the non-response correction (see 3.6), the GREG estimator is a calibration procedure applied to adjust for non-response corrected design weights such that certain benchmark figures of the anchor variables are exactly achieved. According to the BA's employment statistics, the population of establishments and employees is used as a benchmark figure. For the matrix by size classes and economic sectors, the number of size classes is limited to six, and 24 economic sectors are considered. Separate extrapolations are made for Eastern and Western Germany.
The GREG estimator transforms input weights into output weights. Without any further restriction, the output weights can be small (also negative) or large. However, negative weights are useless, and a weight below 1 is hard to interpret theoretically. An upper limit is also reasonable, depending on the sample size. The extrapolation procedure applied here restricts the weights to the range [1, 10,000]. This arrangement can be adjusted within the scope of future extrapolations, for example, to account for different sample sizes.
A full description and the background to its development and implications is given in Brenzel, Hanna; Czepek, Judith; Kiesl, Hans; Kriechel, Ben; Kubis, Alexander; Moczall, Andreas; Rebien, Martina; Röttger, Christof; Szameitat, Jörg; Warning, Anja; Weber, Enzo (2016): Revision of the IAB job vacancy survey * backgrounds, methods and results. (IAB-Forschungsbericht, 04/2016 (en)), Nürnberg.
Main dimensions:
Number of local units and number of employees subject to social security contributions
3.6. Adjustment
Non-response adjustment
Initially, in the main survey (in each fourth quarter), a non-response estimation is made with the added auxiliary variables based on register data. An adjustment weight is calculated for the non-response behaviour. The auxiliary variables (vectors) originate from the BA's administrative data (cf. Section 4.3). The affiliation to an economic sector and an establishment size class are included in the estimation in addition to the added average daily wage and the average age of staff. A binary choice model (logit estimator) calculates the response probability for the given auxiliary variables. The inverse of this response probability of an establishment is the non-response weight (NR) for the main survey in the fourth quarter.
Since the follow-up surveys in the following first, second and third quarters depend on the response behaviour in the main survey, a second-stage non-response adjustment is applied: The first component is determined by the non-response weights of the main survey. The second component is a non-response analysis conducted separately for each following quarter based on the auxiliary variables' information and the main survey's response behaviour. So, on the one hand, a check is made first for non-responses compared to the gross sample (or the total number of establishments with one employee subject to social insurance contributions from BA statistics) and, on the other hand, based on the main survey for non-participations in the quarterly follow-up surveys. Although the willingness to participate is significantly higher than in the main survey, specific response behaviour can also lead to systematic distortions in the follow-up survey.
4.1. Quality assurance
To check data quality, the quarterly subset of job vacancies that are reported to be registered at the Federal Employment Agency are compared with official register data.
4.2. Quality management - assessment
In the survey of Q2 2021, we detected a significant deviation between the number of survey-reported registered vacancies and the number of registered vacancies at the Federal Employment Agency. This divergence led us to an analysis of interviewer effects in the telephone interviews of the JVS survey. The analysis of interviewer effects was possible by merging a unique interviewer identifier to the micro data. Since the interviewed establishments are randomly allocated to the interviewers by the telephone institute, there should be no systematic interviewer effects. However, we detected that a specific interviewer significantly less likely (statistical precision 99.9%) inserted vacancies. We also reviewed voice files of the interviews of the respective interviewer. Our qualitative conclusion is that the respective interviewer, unfortunately, provided an unacceptable quality of interviewing. The respective interviewer was confronted with our accusation, and he/she admitted fraudulent behaviour.
As a consequence, we excluded the respective interviewer's observations from the micro data. Since the interviewer conducted interviews for JVS not only in Q2 2021 but also in Q1 2020, Q2 2020, and Q3 2020, we excluded interviews from all these quarters and revised the JVS data, respectively.
In the course of our rigorous investigation of fraudulent behaviour in the JVS data, we detected a second interviewer with significant interviewer effects. This second interviewer only conducted interviews in Q1 2020. Consequently, we also excluded the respective interviews of this second interviewer.
5.1. Relevance - User Needs
Description of the national users and their main needs
Remarks
National users:
The public via the media (TV, radio, newspapers, magazines): quarterly press release on the development of job vacancies in East and West Germany by sectors
Federal Employment Agency, Federal and State Ministries (esp. labour market and social policy, immigration policy): short term labour demand, labour and skill shortages, strategies against shortages
International users:
European Commission: publication of data by Eurostat, European Job Vacancy Monitor
European Central Bank and OECD: short term labour demand
Use by national and international scientific researchers:
economic research: matching and mismatch, recruitment processes
sociological research: the effect of labour market reforms, the role of networks in the recruitment process
survey methodology: survey design, non-response-analysis, multiple imputation methods, small area estimation
5.2. Relevance - User Satisfaction
Extent to which the needs of national users are satisfied (voluntary)
Remarks
The IAB Job Vacancy Survey measures the volume of unfilled jobs in the surveyed quarter. It covers the German economy with all private enterprises, associations and public institutions as far as they employ at least one registered employee[1]. Private households and exterritorial institutions are excluded.
Job vacancies include all posts for which the company actively searches qualified staff at external labour markets. They include all types of labour contracts for a dependent job, irrespective of social security registration, duration of employment or occupational status. Job vacancies include posts for immediate recruitment and later recruitment, as long as the company is actively searching.
The survey measures the number of occupied posts including all types of employment, such as dependent employment, self-employment, civil service and marginal employment.
The 4th quarter survey 2010 is the first survey which fully applies NACE Rev. 2 and is therefore able to provide data in the 19 sector breakdown.
As the sample for the 4th quarter 2009 was established on the basis of NACE Rev. 1.1, the representation of NACE Rev. 2 is not available. The reclassification of micro-data according to NACE Rev. 2 was tested. The results were insufficient as particular sectors, such as information and communication or real estate activities, were not adequately represented in the former sample.
[1] Registered by public social insurance.
5.3. Completeness
Description of missing variables and missing breakdowns of the variables
Report progress on the implementation measures regarding quarterly job vacancies statistics of Regulation (EC) No 453/2008, including :
a detailed plan and timetable for completing implementation
a summary of the remaining deviations from EU concepts
Regulation is fulfilled.
Regulation is fulfilled.
5.3.1. Data completeness - rate
Not applicable.
6.1. Accuracy - overall
Not applicable.
6.2. Sampling error
See below.
6.2.1. Sampling error - indicators
Coefficient of variation (taking into account the sampling design) or estimated sampling error for the number of job vacancies (see guidelines). See also Table 11.
Coefficient of variation for the whole economy (%)
Year
Quarter
CoV full sample
CoV 10+ employees
2024
1
3%
4%
2024
2
3%
3%
2024
3
3%
3%
2024
4
3%
3%
For coeffients of variation for each NACE by quarter, see Tables 7-10.
Information on variables with non-negligible measurement and processing errors
Information on main sources of (non-negligible) measurement and processing errors and, if available, on methods applied for correction
Estimation bias: An assessment of the non-sampling errors, in terms of the absolute number of vacant posts, for the total number of job vacancies and, where possible, for aggregation level of NACE Rev. 2 specified in Annex 1 to this Regulation and size classes (1-9, 10 + employees).
Remarks
The strong diversity in the number of vacancies creates a serious sensitivity of the calibration method to extreme values. This is due to the fact that vacancy information shows a high degree of variation at the micro-level. One third of the business units offer not more than one vacancy at the time of observation, and not more than 5 % offer 25 vacancies or more.
The new weighting procedure is constructed without an adjustment to registered vacancies. The number of employees subject to social security contributions and the number of establishments are used as anchor variables. This method allows considering non-response corrections and the direct deduction of key figures to assess the quality of the extrapolation and the validity of results which allow statements regarding the statistic validity of the survey. The GREG procedure is more efficient in terms of methodology than the previously used procedure. It significantly improves the quality of the extrapolation.
see 3.5
see 3.6
6.3.1. Coverage error
Description of any difference between the reference population and the study population
Description of classification errors
Description of any difference between the reference dates and the reference quarter
Any other relevant information
The sampling frame (the business register maintained by the German Federal Employment Agency) consists of all business units that have at least one employee covered by the public social insurance system. Any other business units are not part of the sampling frame, and therefore the estimates might slightly be biased downwards as they exclude companies with non-registered employees only. A similar bias might occur, because the sampling frame is updated with a time lag of several months. Very new business units are thus excluded from the sampling frame; however, the frame might contain units that ceased to exist after the latest frame update. As the estimates are calibrated to the (estimated) current number of business units in the population during the period of observation, the effects of such biases are largely corrected.
See first column
The survey is conducted as quarterly average.
See first column
6.3.1.1. Over-coverage - rate
Not applicable.
6.3.1.2. Common units - proportion
Not applicable.
6.3.2. Measurement error
See 6.3.
6.3.3. Non response error
See 3.6.
6.3.3.1. Unit non-response - rate
Unit non-response rate (see also Tables 2 - 5)
1st quarter 2024
2nd quarter 2024
3rd quarter 2024
4th quarter 2024
30.3 %
31.1 %
34.0 %
88.6 %
To cope with the low response rates and to minimize non-response bias as far as possible, the calibration procedure described in the section on Data compilation (weighting procedure) is used for estimation.
6.3.3.2. Item non-response - rate
Not applicable.
6.3.4. Processing error
See 6.3. Non-sampling error.
6.3.4.1. Imputation - rate
Item imputation rate and methods and, where possible, the effect of imputation on the estimates for the variables transmitted
No imputation procedures for item non-response concerning the number of job vacancies were applied.
The occupied posts are included in the survey each quarter. However, in very few cases this variable is missing due to item non-response of the respective establishments. In such cases we impute the number of employees from the previous quarter. If it is again not available in the previous quarter, we go back up to the previous Q4 wave, where we have the number of employees for each firm in the data.
6.3.5. Model assumption error
If modelling is used, include a description of the models used. Particular emphasis should be given to models for imputation or grossing-up to correct for unit non-response.
A new extrapolation procedure based on the Generalized regression estimator (GREG) was developed in close collaboration between the IAB, the Economix Research and Consulting survey research institute and the Regensburg University of Applied Sciences. This estimator is widely used in modern survey research (cf. Särndal et al. 1992; Deville/Särndal 1992) and is also recommended by the European Commission, for instance (cf. European Commission 2002). See points 3.5 and 3.6
6.4. Seasonal adjustment
Brief description of seasonal adjustment procedures, in particular with regard to the European Statistical System guidelines on seasonal adjustment which have been endorsed and supported by the SPC.
Since the fourth quarter of 2014, the time series collected is sufficiently long to allow for seasonal adjustment. The seasonally adjusted time series is calculated using the jDemetra+ Software provided by Eurostat. The preferred adjustment follows the TRAMO/SEATS procedure and is calculated over the entire series for each quarter. It is done on the level of aggregated series provided to Eurostat[2] for all local units and local units with at least ten employees.
From the fourth quarter of 2016 onwards, there are regular revisions of past observations and their seasonal adjustment. This adjustment will be based on updates of the administrative data.
[2] Given that Eurostat has aggregated the individual series for EU-wide statistics, the earlier method to provide series by 19 sectors with seasonal adjustment has not been continued. Eurostat requested to provide the aggregated series with seasonal adjustment.
Provide a revision history, including the revisions in the published number of job vacancies and a summary of the reasons for the revisions.
In July/August 2026, the vacancies and occupied posts of 2024 will be (regularly) revised according to the final administrative employment figures of the Federal Employment Agency.
6.6.1. Data revision - average size
Not applicable.
7.1. Timeliness
See below.
7.1.1. Time lag - first result
Information on the time span between the release of data at national level and the reference period of the data.
Data are sent to Eurostat within 45 days after the end of the reference quarter.
7.1.2. Time lag - final result
Not applicable.
7.2. Punctuality
See below.
7.2.1. Punctuality - delivery and publication
Deadlines for the respondents to reply, also covering recalls and follow-ups (see Table 12)
Period of the fieldwork
Period of data processing
Dates of publication of first results
Remarks
1st quarter 2024
2nd quarter 2024
3rd quarter 2024
4th quarter 2024
Fieldwork
01 January - 30 March 2024
01 April - 29 June 2024
01 July - 30 September 2024
28 September - 31 December 2024
Assembly of data and quality control
15 April - 30 April 2024
15 July - 31 July 2024
15 October - 13 October 2024
12 January - 05 February 2025
Calibration
06 May 2024
05 August 2024
05 November 2024
11 February 2025
Reporting, delivery of data
13 May 2024
09 August 2024
11 November 2024
30 March 2025
See table in the first column
See table in the first column
Data are available within six weeks after the end of the quarter.
Information on differences between national and European concepts, and — to the extent possible — their effects on the estimation.
The revised data are fully consistent with Eurostat standards.
8.1.1. Asymmetry for mirror flow statistics - coefficient
Not applicable.
8.2. Comparability - over time
Information on changes in definitions, coverage and methods in any two consecutive quarters, and their effects on the estimation.
The time series has been weighted using the new weighting method for provided data. This produces a consistent series concerning the weighting procedure since 2010.
The weighting procedure is described in Brenzel et al. (2016).[3]
Comparisons of data on the number of vacant jobs from other relevant sources when available, in total and broken down by NACE at section level when relevant, and reasons if the values differ considerably.
There are no other administrative sources or other representative surveys on the population of job vacancies in Germany. The administrative data from the Federal Employment Agency only cover the number of registered vacancies, which covers about half of the total number of vacancies.
Attached you can find a comparison of occupied posts between data of the JVS and the LFS.
Dissemination scheme, including to whom the results are sent
Periodicity of national publication
References for publications of core results, including those with commentary in the form of text, graphs, maps, etc.
Information on what results, if any, are sent to reporting units included in the sample
JVS data are disseminated online by the IAB in German language: Iab-stellenerhebung.
Aggregated data are published online
Publications that apply the IAB-Job Vacancy Survey can be accessed through: Iab.
Not applicable.
9.3. Dissemination format - online database
Not applicable.
9.3.1. Data tables - consultations
Not applicable.
9.4. Dissemination format - microdata access
Not applicable.
9.5. Dissemination format - other
Not applicable.
9.6. Documentation on methodology
Not applicable.
9.7. Quality management - documentation
Description of and references for metadata provided
References for core methodological documents relating to the statistics provided
Description of main actions carried out by the national statistical services to inform users about the data
No specific metadata are provided. The quarterly press release contains all important information regarding meta data changes.
Kettner, Anja, Vogler-Ludwig, Kurt (2010): The German job vacancy survey. An overview. In: European Commission, Eurostat (Ed.): First and second International Workshops on Methodologies for Job Vacancy Statistics. Proceedings, Eurostat Methodologies and working papers, Luxemburg, S. 7-17.
Bossler, Mario; Gürtzgen, Nicole; Kubis, Alexander; Küfner, Benjamin; Lochner, Benjamin (2020): The IAB Job Vacancy Survey: design and research potential. In: Journal for labour market research, Vol. 54, No. 1, Art. 13.
Bossler, Mario; Kubis, Alexander; Küfner, Benjamin; Popp, Martin (2021): The IAB Job Vacancy Survey: Establishment survey on labour demand and recruitment processes, waves 2000 to 2018 and subsequent quarters 2006 to 2019. (FDZ-Datenreport, 09/2021 (en)), Nürnberg, 19 S.
Data are provided with a quarterly press release. New results and visible trends are intensively discussed by the national media on base of this release.
9.7.1. Metadata completeness - rate
Not applicable.
9.7.2. Metadata - consultations
Not applicable.
Not applicable.
11.1. Confidentiality - policy
Not applicable.
11.2. Confidentiality - data treatment
Disclosure rules: Brief description of when data have to be deleted for reasons of confidentiality
No data are deleted for reasons of confidentiality.
A 'job vacancy' is defined as a paid post that is newly created, unoccupied, or about to become vacant:
for which the employer is taking active steps and is prepared to take further steps to find a suitable candidate from outside the enterprise concerned; and
which the employer intends to fill either immediately or within a specific period of time.
‘Active steps to find a suitable candidate’ include:
notifying the job vacancy to the public employment services,
contacting a private employment agency/head hunters,
advertising the vacancy in the media (for example internet, newspapers, magazines),
advertising the vacancy on a public notice board,
approaching, interviewing or selecting possible candidates/potential recruits directly,
approaching employees and/or personal contacts,
using internships.
'Specific period of time’ refers to the maximum time the vacancy is open and intended to be filled. That period shall be unlimited; all vacancies for which active steps are continuing on the reference date shall be reported.
An 'occupied post’ means a paid post within the organisation to which an employee has been assigned.
Local unit.
All establishments with at least one or more employees.
All establishments located in the territory of Germany.
Not Applicable
Not applicable.
Not Applicable
Brief description of the weighting method
Weighting dimensions
In 2015, a new extrapolation procedure based on the Generalised Regression Estimator (GREG) was developed in close collaboration between the IAB, the Economix Research and Consulting survey research institute and the Regensburg University of Applied Sciences. This estimator is widely used in modern survey research (cf. Särndal et al. 1992; Deville/Särndal 1992) and is also recommended by the European Commission (cf. European Commission 2002).
The GREG procedure continues the basic ideas of the extrapolation method used in the past. However, the adjustment to registered vacancies as an anchoring variable has been discontinued. Now, the number of employees subject to social security contributions and the number of establishments are used as anchor variables. Note that we base these numbers on the most recent available data, which usually lags by 2-3 quarters. A forecast of employment development is used to adjust the anchor variable to the current employment levels in the German economy.
The method allows the implementation of different kinds of weights. It allows considering non-response corrections and the direct analytical calculation of key statistics to assess the quality of the extrapolation and the validity of results, which allows statements regarding the survey's statistical validity. The GREG procedure is more efficient in methodology than the previously used procedure. It significantly improves the quality of the extrapolation.
GREG Implementation:
After the non-response correction (see 3.6), the GREG estimator is a calibration procedure applied to adjust for non-response corrected design weights such that certain benchmark figures of the anchor variables are exactly achieved. According to the BA's employment statistics, the population of establishments and employees is used as a benchmark figure. For the matrix by size classes and economic sectors, the number of size classes is limited to six, and 24 economic sectors are considered. Separate extrapolations are made for Eastern and Western Germany.
The GREG estimator transforms input weights into output weights. Without any further restriction, the output weights can be small (also negative) or large. However, negative weights are useless, and a weight below 1 is hard to interpret theoretically. An upper limit is also reasonable, depending on the sample size. The extrapolation procedure applied here restricts the weights to the range [1, 10,000]. This arrangement can be adjusted within the scope of future extrapolations, for example, to account for different sample sizes.
A full description and the background to its development and implications is given in Brenzel, Hanna; Czepek, Judith; Kiesl, Hans; Kriechel, Ben; Kubis, Alexander; Moczall, Andreas; Rebien, Martina; Röttger, Christof; Szameitat, Jörg; Warning, Anja; Weber, Enzo (2016): Revision of the IAB job vacancy survey * backgrounds, methods and results. (IAB-Forschungsbericht, 04/2016 (en)), Nürnberg.
Main dimensions:
Number of local units and number of employees subject to social security contributions
Identification of the source of the data
IAB-Job Vacancy Survey
Coverage
Geographical
The whole territory of the country
NACE
NACE sections A-S
Enterprise size
Establishments with one and more employees
Definition of the statistical unit
Local unit
Sampling design
Base used for the sample
For the postal survey in the fourth quarter of 2024, a sample of about 140,000 local units is drawn randomly from the official business register of the Federal Employment Agency. The sampling ratio is 6.6 % of the population of local units with registered employees in Germany (see Table 1).
The sample is stratified by:
2 regions (East and West Germany)
24 sectors (NACE Rev. 2 since IV/2010)
7 size classes
The number of local units in every cell of the sample matrix is chosen regarding the participation rates in previous surveys to ensure a sufficient number of responses. The probability of having an open vacancy is especially low in very small units. On the other hand, very small units account for about 30 per cent of all vacancies. For this reason, the absolute sample size is larger in smaller-sized classes (see Table 1).
The net sample size of the postal survey in the fourth quarter of 2025 was 15,905, thereof 11,311 in the sample for West Germany and 4,594 in the sample for East Germany (Table 2).
From the 4th quarter of 2023, the invitation for participating organisations was changed and now focuses more on online participation. After the initial invitation to participate online, one reminder with the paper questionnaire enclosed was sent to the companies that had not yet responded. The new procedure contributed to a higher response rate compared to previous years; the share of online responses has also increased.
In addition to the 4th quarter, this new mode was also used for the first three quarters of the following year.
Due to a high response rate in the 4th quarter of 2023, the samples for the follow-up surveys in the first quarter of 2024 consisted of around 18,000 local units. These are subsamples of the previous year's main survey (Tables 3 to 5). The net sample size in the first three quarters of 2024 averaged over 12,000.
The four surveys thereby formed a quasi-panel from the 4th quarter to the 3rd quarter of the following year. In the observation period 2023/24, 44.3 % of participants took part in all four surveys, 16.8 % in at least three, and 13.9 % in at least two of the surveys (Table 6). 25 % only participated in the main survey.
The response rate for the main survey in 2024 was 11.4%, and for the follow-up surveys, it ranged from 69.7% in the first quarter to 66.1% in the third quarter of 2024.
Sampling design
The number of local units in every cell of the sample matrix is chosen regarding the participation rates in previous surveys to ensure a sufficient number of responses. The probability to have an open vacancy is especially low in very small units. On the other hand, very small units’ account for about 30 percent of all vacancies. For this reason, the sample size is larger in small size classes (Table 1).
Retention/renewal of sampling units
The sample is renewed once a year in the fourth quarter.
Sample size
Gross sample size:
Q1 2024: 18,268 local units
Q2 2024: 18,190 local units
Q3 2024: 17,865 local units
Q4 2024: 139,247 local units
Net sample size:
Q1 2024: 12,725 local units
Q2 2024: 12,527 local units
Q3 2024: 11,783 local units
Q4 2024: 15,905 local units
Stratification
The samples for Eastern and Western Germany are stratified by
24 sectors
7 size classes
Other sources
Maintenance agency
The survey is conducted under the responsibility of the Institute for Employment Research (IAB) at the German Federal Employment Agency. On behalf of the IAB, the 2024 survey was conducted by Economics & Data ED23 GmbH in Munich.
Information on differences between national and European concepts, and — to the extent possible — their effects on the estimation.
The revised data are fully consistent with Eurostat standards.
Information on changes in definitions, coverage and methods in any two consecutive quarters, and their effects on the estimation.
The time series has been weighted using the new weighting method for provided data. This produces a consistent series concerning the weighting procedure since 2010.
The weighting procedure is described in Brenzel et al. (2016).[3]
