This website uses cookies and the Web Analysis tool Piwik. This helps us to provide you with a good user experience and to improve our website. If you surf through our sites, you hereby agree. Here you will learn more about the use of your data and possibilities for contradiction.

Quality of data collection in the job-centres

Federal agency receives information from the German Federal Court of Audit
 
18 Mar 2019 | Press Release No. 11

What has the Federal Court of Audit examined?

The Federal Court of Audit (BRH) has examined, in its preliminary report, whether the data collected, inter alia, for surveying for unemployment in the job-centres, are correct. It noticed deficiencies in the process.

The Federal Court of Audit notes, for example, that some "unemployed" were identified as "job-seekers" or vice versa. Or, clients who were identified as "job-seekers" during integration or reintegration were initially not returned to "unemployed" status, after the measure ended.

The Federal Court of Audit does not assume in its draft report that this could be deliberate manipulation or a systematic error. The German Federal Employment Agency takes the advice of the Federal Court of Audit seriously, and as an opportunity to continue working on the quality of data collection in the unemployment statistics and to further improve service in the job-centres.

Why have there been errors detected?

All data from job-centre clients, including their status, are recorded in the interview and entered manually. Most of the status-changes are then automatic.

Most errors are where entries have to be changed manually. This can be the case after a measure that lasts longer than six weeks and thus legally interrupts the "unemployed" status. Automatically generated reminders remind the job-placement agents that a qualification measure ends and the customer must be invited to an interview. Among other things, the agent must manually update the status during this interview.

The fact that the status of a service recipient in the data system is incorrectly recorded can have various reasons; it can simply be due to an incorrect entry. But there are also situations in which everything is done right and the status is still inaccurate for the time being. Example: A customer is invited to an appointment after the measure ended, during which the further course of action is to be discussed. It also sets the current status, for example, from "looking for work" to "unemployed". If the client is unable to attend the appointment, the status cannot be updated. It is required by law that an unemployment report can only be made in person. As long as this does not happen, the agent must not change the status on his / her own.

What is the BA doing to prevent erroneous data entry?

Additional test programs, which also may indicate discrepancies in status, should be made mandatory and used even more reliably, in the future.

In the future, the job-centres will work even more intensively in client care to ensure the smoothest possible support. For this, the BA will invest even more in ensuring the quality of the mediators' work, for example with training courses. Only then can the status of the customers be entered as correctly as possible into the systems.

What does this mean for statistics about the labour market?

The labour market statistics of the German Federal Employment Agency are based on a full survey. All persons with "unemployed" status in the IT systems at the deadline, will be counted. However, a statistical margin of error must be taken into account, with all full surveys. The source data is aggregated into a statistic and systematically validated, using various methods. For example, at least once a month, it is technically and substantively checked whether the data input is complete, the data is consistent in comparison to other statistics, and whether movement- and inventory data match.

As part of the official statistics, the statistics of the BA is obligated to follow the principle of quality and works here and in the future consistently on further improvements. Even though the absolute numbers can never be correct to the last person, and there may be a slightly lower number of unemployed, the BA statistics correctly reflect the level and the changes and trends.

For a comprehensive labour market observation, besides unemployment, we also draw other variables. These are, the employment subject to social security contributions and the reported job openings. These are not affected by the facts mentioned in the preliminary report.

Therefore, nothing changes in the (previous) statements of the BA on labour market development. Unemployment continues to follow a declining course, as does the unemployment rate, which is currently the lowest since the reunification.