Last update: 4 March 2021

Change in the methodology of the data collection system for retail trade statistics and in the publication of data.

Publication of retail sales data

Publication of retail and catering data will change with effect from 5 March 2019: current data will be published based on a new methodology, in which data collection is supplemented with data from administrative sources. Based on the new methodology, we can generate these data based on more comprehensive population data compared to previous practice, resulting in reduced sampling error. In preparation for the use of online cash register data, HCSO examined the impact of the new data source on the 'whitening' of the economy. On this basis, as a result of the new methodology using online cash register data, data provide a more reliable picture of retail and catering sales than the previous practice, which was rather based on sampling

Data can be accessed on the HCSO website at the following links: Tables (STADAT) ? Time series of annual data ? Internal trade; Sales of retail shops by type of shop (2000?).

In order to ensure data comparability over time, sales data were backcasted until January 2015 and volume indices until January 2016, which are also available in the tables and database. The structure of the published tables has not changed compared to the previous ones, they contain the usual monthly and quarterly information. Data collected based on the methodology used before 2019 will remain available to users by clicking on the following link: Tables include data until December 2018, but please note that the values shown here are not fully comparable to the data calculated on the basis of the new methodology.

Under EU Regulation 2019/2152 on European Business Statistics, repealing 10 legal acts in the field of business statistics and EU Regulation 2020/1197 on implementation thereof, from 2021, retail sales should include retail sales via stalls and markets1 (TE┴OR?08 47.8) and other retail sale not in stores, stalls or markets (47.99), within the non-food group. For comparability, data on market and other non-store retail sales are available at current prices from 2019, while volume index data are available from 2020 onwards. This regulatory change increases current price turnover by 0.6-0.8%, while it has no significant effect on the volume index2.

You can find detailed information about the renewed methodology below:

Estimating retail shop data according to the new methodology

Retail trade and catering turnover and volume index is based on a survey launched by electronic questionnaires – extended with administrative data –, where the respondent is the enterprise and the observation unit is the shop. The target population is provided by the Retail Trade Register (KERREG) which is updated from shop registers biannually received from municipalities. The population is divided into a full scope and a surveyed subpopulation.

Based on NGM Decree 48/2013 (hereinafter referred to as Decree), the use of online cash registers (OPGs) was introduced in retail trade Using these data, HCSO has increased the coverage of data providers while reducing respondent burdens by approximately 80%.

The majority of enterprises surveyed receive simplified questionnaire and a smaller part full questionnaire:

the simplified? questionnaire is for those enterprises whose shops (all of them) are under the Decree and do not engage in any other activity, these enterprises have to report the number of their shops, and the sum of the turnover of the retail trade, catering, online orders and home delivery sales;

and the original ?full? questionnaire for the rest of the full scoped enterprises and for the population not covered by the Decree and surveyed by simple random sampling.

It is still necessary to maintain a simplified data collection via questionnaire, because:

there are enterprises/shops not under the Decree and operating outside the OPG system;

there is no other source of data about online orders and home deliveries;

in some cases (operation system error, other activity, etc.) there is too much difference between the two data sources.

Data are estimated in the following way:

In case of full scope observed enterprises using a simplified questionnaire, questionnaire data are preferred and the quotient of these data and data from the administrative sources are used when producing detailed statistics (e.g., by activity or region).

For the estimation of the non full scope part we use a monthly correction for the cash register data in order to compensate late response due to technical reasons.

The new data source will not affect the estimation of tobacco, mail order and internet retail sales. The estimation for the enterprises/shops not covered by the Decree remains unchanged.

In case of OPG data, exceptionally large items are handled as follows: NAV files including items above a specified threshold and sent monthly by NAV are aggregated according to cash registers during processing. Outlier values are kept (accepted) in case of full-scope observed enterprises and automatically deleted (subtracted) from total sales in case of all other shops.

Missing data are not imputed in the administrative data source. In the full scope part we impute with the data for the same period of the previous year if they exist. If there is no such data but previous period data is available, the product of this and the average dynamics (the rate of the average of the period and the previous period) will be the imputed data. The same imputation will be used in the case of the sampled units. If no prior period data is available we impute with a corrected average value.

Effect of methodological change on the estimation of retail sales data by major commodity groups

For the estimation of retail sales data by major commodity groups, we also use the aggregated data of retail shops produced according to the new methodology. For this reason, in order to ensure consistency and data comparability overt time, commodity group data were also backcasted until the beginning of 2015. For the same reason, from January 2019 data also include market and other non-store retail sales.

The methodology used for the estimation of retail sales data by major commodity groups remains unchanged in other respects and is still based on a sampling survey, while data of the Tobacco commodity group is based on comprehensive administrative tobacco sales data provided by the National Tobacco Sales Non-profit Ltd. In order to estimate the turnover of each commodity group, the aggregated turnover of the retail and automotive retail stores is divided by the percentage distribution obtained from the sampling survey. In case of the Tobacco commodity group we publish the received aggregated administrative data unchanged.