Statistical processing

Source data

Danish foreign affiliates statistics is a census survey (not a sample survey), which means that the statistics is intended to be based on information from all Danish companies that own or control affiliates abroad. At the same time this means that a grossing up of the figures is not carried out.

The survey is a mandatory questionnaire based survey.

Sources for background information is a number of internal sources notably the Business Register and external sources, notably the Foreign Direct Investment Register of the Central Bank, Annual Company Reports and commercial data sources. In that way all enterprises in Denmark with controlling influence in foreign affiliates will contribute to this statistics.

Background information regarding the Danish companies are based on internal sources in Statistics Denmark, primarily the Business Register.

Frequency of data collection

Information on Danish foreign affiliates is collected annually.

Data collection

Danish foreign affiliates is an questionnaire based survey. The questionnaire are mainly answered online via https://www.dst.dk/virk, however some mainly large Danish company use alternative methods for reporting.

In order to ensure a high response-rate, written reminders are sent by the postal system, and also, if possible, by email.

Data validation

The received data for Danish foreign affiliates are validated both during the collection of data phase, but mainly when data collection is completed.

Several data validation procedures which are different in nature are conducted on the transmitted data in order to ensure a high data quality. Some variables are sought completely fulfilled, e.g. the country of the affiliate and the activity conducted in this. The reason is both to ensure high data quality but also in order to determine relevance (meaning is the affiliate a part of the scope of the survey).

The data on Danish foreign affiliates are validated and corrected on micro level, meaning the individual affiliate. Next the data are validated on reporting unit level, meaning the Danish company. And finally the data are validated on macro level, this means on the aggregated data broken down on the affiliates main activity and/or geographical break down. The data validation is prioritized, and errors/developments that have a huge impact on the statistics are validated first, thus meaning that a selective and prioritised data validation is carried out.

With micro-level data-treatment complete, the data is validated on a more aggregated level, e.g. on main activity breakdown of the combined affiliates, on the geographical breakdown of the combined affiliates and also on reporting unit level. The purpose is here to validate the developments from one year to the next, in order to catch major developments which could indicate possible errors.

The data validation on micro level is undertaken in order to correct actual errors, whereas the data validation on macro level aims at trying to understand the data and to find sound explanations on developments found between survey-years. This is carried out in close contact with the reporting unit.

Insufficient or faulty data submissions from large and thereby significant units (based on the previous year data transmission) are sought made complete and corrected, primarily by contacting the reporting unit.

In the case of missing information after this, values are imputated - primarily by using available register or annual report information, alternatively the data reported in the previous year. In that regard it should be mentioned that it is mainly on data transmissions containing zero-values.

Data compilation

Danish foreign affiliates statistics is a census survey (not a sample survey), which means that the statistic is intended to be based on information from all companies in Denmark that own or control affiliates abroad. At the same time this means that a grossing up of the figures is not carried out. Furthermore, and due to the generally high response-rate, imputation of missing units (unit non-response) is only carried out to an insignificant extent, primarily by using information available in registers or annual reports, alternatively in the data submitted the previous year.

Adjustment

No other correction of data are carried through, besides the corrections mentioned under data validation and data compilation.