Accuracy and reliability

Overall accuracy

At the first publication of a given income year, approximately 99.0 pct. of the people who are liable to taxation has been assessed. After the income taxes have been grossed up, the total income tax equivalates approximately 99.4 pct. At the second publication of a given income year the number of finalized assessment notices increases such that approximately 99.6 pct. of people who are liable to taxation have been assessed. After the income taxes have been grossed up, the total income tax equivalates approximately 99.9 pct. At the third and final publication of a year the assessment rate for people who are liable to taxation increases to 99.9 pct. After the income taxes have been grossed up, the total income tax equivalates approximately 99.96 pct.

The reason why the income taxes will not amount to 100 pct. is that the Danish Tax Agency updates their data on a continuous basis as more people who are liable to taxation are assessed and any changes to already finalized assessment notices are incorporated into the data.

Sampling error

Not relevant for these statistics.

Non-sampling error

The process of finalizing all tax assessment notices is complex. Some notices take longer to finalize than others. In the first dissemination of the statistics, a percentage of the population liable to taxation in Denmark will be missing in the data set because their tax assessment notices have yet to be finalized. Consequently, it is necessary to gross up the taxes in the data set in order to describe the income taxes for the entire population, who are liable to taxation in Denmark. The data is grossed up by multiplying a coefficient to the already reported taxes. The coefficient is calculated as the ratio of persons who have had their tax assessment finalized compared to the total number of personals liable to taxation. This procedure is implemented in both preliminary and final versions of the statistics.

Non-response error will be greatest at the first publication. Non-response error emerges when data does not include the entire population and it occurs because of the following conditions. Some people have extended deadlines for reporting to the Tax Authorities, some people choose not to report to the tax authorities, some cases are complex and take longer to complete, and finally, there might be corrections to already finalized tax assessment notices. Therefore, the coefficient used to gross up the data will be largest at the first publication. As more data are assessed, the need to gross up the data will be smaller. At the final version of an income year it is, however, still necessary to gross up the data since the data might still contain non-response errors.

The Danish tax authorities continually implement corrections to data, which minimizes measurement errors. Measurement errors occurs when the statistics does not measure correctly. Experience from Statistics Denmark has shown, however, that beyond 1½ year after the end of the income year, only minor changes will happen to the final assessments. Some tax cases are not concluded until several years after the end of the income year. The tax authorities will continue to implement these corrections to specific datasets concerning changes, and as a result of this the final assessment will never be completely "final".

For some preliminary tax concepts, it is necessary to correct for potential bias. The reason for this is that the tax concepts are underestimated compared to the final value. The bias corrections are based on previous year’s difference between preliminary and final values. Thus, the corrections are removed when a year is finalized.

The statistics cover almost every person who is liable to taxation in Denmark, which is why coverage error is minimal for these statistics. The reason for the minimal coverage error is that data changes can be allocated back to the year the data change actually belongs to. This means that if a notice is not finalized until 2021, even though it pertains to income taxes in 2019, the notice will be allocated to the income taxes in 2019 and not affect the income taxes in 2021. Due to Statistics Denmark's revision practice corrections are not carried further back than three years. The years beyond that are considered final.

Quality management

Statistics Denmark follows the recommendations on organisation and management of quality given in the Code of Practice for European Statistics (CoP) and the implementation guidelines given in the Quality Assurance Framework of the European Statistical System (QAF). A Working Group on Quality and a central quality assurance function have been established to continuously carry through control of products and processes.

Quality assurance

Statistics Denmark follows the principles in the Code of Practice for European Statistics (CoP) and uses the Quality Assurance Framework of the European Statistical System (QAF) for the implementation of the principles. This involves continuous decentralized and central control of products and processes based on documentation following international standards. The central quality assurance function reports to the Working Group on Quality. Reports include suggestions for improvement that are assessed, decided and subsequently implemented.

Quality assessment

The overall quality of the statistics is considered very good because of the comprehensive data foundation available for the production of personal income taxes. Data is collected from the Danish Tax Agency's administrative registers, which is updated on a continuous basis as more people are assessed and any changes to already finalized assessment notices are incorporated into the data. Already at the first publication of the statistics approximately 99.0 of the people who are liable to taxation have been assessed. After the grossing up to the total population, the total income tax constitutes approximately 99.4 pct. At the third and final publication of a year the assessment rate for people who are liable to taxation increases to 99.9 pct. After the grossing up, the total income tax constitutes approximately 99.96 pct.

Statistics Denmark processes the data from the Danish Tax Agency by grossing up the data to ensure that every person liable to taxation in Denmark is represented in the data set. In addition, it is also necessary to correct for potential bias, since some tax concepts are underestimated in the preliminary version compared to the final value. The purpose of this procedure is to minimize non-sampling error in the data. Subsequently, data is validated by comparing the new income level to the level of the previous year or publication. The comparison takes both business tendencies and possible tax rate changes into account.

The first version of preliminary personal income taxes is published in November after the end of the income year. There are normally only minor revisions between the provisional and final numbers. The revision between the first and the final publication is 0.6 pct. of the total income tax, whereas the revision between the second and final publication is less than 0.1 pct. The final income taxes are published in September, three years after the end of the income year in question. The revision frequency is consistent with national accounts.

The statistics are part of the general economic debate and often gets a lot of attention in the media and among other professional users.

Data revision - policy

Statistics Denmark revises published figures in accordance with the Revision Policy for Statistics Denmark. The common procedures and principles of the Revision Policy are for some statistics supplemented by a specific revision practice.

Data revision practice

Some tax cases are not concluded until several years after the end of the income year. The Danish tax authorities will thus continually receive corrections to data. The corrections are added to specific datasets concerning changes, and as a result of this the final assessment will never be completely “final”. Experience has shown, however, that beyond 1½ year after the end of the income year, only minor changes will happen to the final assessments. Statistics Denmark thus only makes corrections concerning the latest three years while earlier years are considered as final.