Statistical processing
Contact info
Labour and Income, Social StatisticsJarl Quitzau
+45 23 42 35 03
Get as PDF
Various sources are used for the register. Some components are delivered directly from the Tax administration. Regarding pension wealth assumptions are made to distribute bonuses. Finally the valuations of real estate, cars and unquoted shares have been imputed. All the data are compiled using anonymised personal-identifiers.
The first version of the 2023 release of pension assets is partially based on a projection. For approximately 6 percent of the assets, it has been necessary to project 2022 data.
Source data
Data is collected from a variety of sources. Housing evaluation stems from the Danish property Assessment Authority but are adjusted to market values using sales statistics and the register on buildings and dwellings (BBR). Most data on financial assets and debt comes from the tax authorities. The value of cars is assessed by using used car sales prices from AutoBranchen Danmark and ownership of the cars are based on the car register. The pension data are collected from each of the 110 pension funds in close cooperation with the Danish National Bank and the tax authorities. The municipalities (mainly KMD) delivers data on postponed payments of real estate taxes. The Danish Debt Collection Agency (Gældsstyrelsen) delivers data on public debt. The assessment of the value of unquoted shares are done by Statistics Denmark in cooperation with the National Bank. It's mainly based on accounting statistics. The unquoted shares are distributed via the registers on ownership of businesses. LD delivers data on vacation funds.
Frequency of data collection
Annual.
Data collection
Most data collected from public registers and they are delivered to Statistics Denmark via the CEMOS-system. All data contains personal identification numbers etc. that are used to link the various inputs and to the central personal register. The personal identification numbers are replaced by random identifiers and only information relevant to the production of the statistics enters the data processing phase.
Data validation
The data on pensions is validated by Statistics Denmark and The Danish National Bank. Data is validated on both the micro level and the macro level where the data form each pension provider is compared to aggregated accounting data from The Danish Financial Supervisory Authority. Most other data coming from the tax system are validated by the tax authorities. Data is systematically checked for new codes. Prior to publication, the annual changes to the wealth components are evaluated and compared to secondary sources, where these are available.
Data compilation
The tax authorities real estate estimations are adjusted to better reflect actual sales prices. This is done within strata on region, size of the dwelling and they type of dwelling. The value of cars are estimated using used car sales prices. Prices are in the model assumed to decrease exponentially. The price does not include the surplus of the car dealership. The Scrap Premium are assumed to be the minimum price. Pension holders may have several types of pensions within the same contract. However pension bonuses are only reported at the contract level. Statistics Denmark distributes the bonuses proportionally to the value of the individual pension schemes within the same contract. In pension wealth statistics the company is split by type based on an annual report from the Danish Financial Supervisory Board. The value of civil servant pensions are estimated based on assumptions on current wage, seniority, expected retirement age, interests and expected remaining lifetime. The value of unquoted stocks is estimated based on the equity and adjusted to market value using the relationship between stock prices and equity in noted companies - following recommendation from the OECD and Eurostat. The owners of the stocks are found using administrative registers on the real owners of the businesses. Owners with less than 25 pct. dos not have to be registered - Some shares can thus not be distributed on persons. In total 87 per cent of the values has been distributed on persons in 2021.
Data from the various sources are merged with an anonymized personal identifier. For publications secondary sources on geography, age, income levels, socio economic status etc. are used.
Adjustment
Not relevant for these statistics.