Index of wholesale and retail sales

Turnover includes sales income from goods and services, transport and emulsion cost which is invoiced to the customer, as well as rents, commission fees and royalties. Financial revenues and value added taxes are not included.

Standard Industrial Classification (SIC2007).

Name: Index of wholesale and retail sales
Topic: Wholesale and retail trade and service activities

Division for Business Cycle Statistics

No geographical breakdown available. National level only.

Monthly. The statistics are normally published between the 26th and 30th day of the following month.

The statistics are reported to Eurostat at the time of publication in Norway.

Collected and revised data are stored securely by Statistics Norway in compliance with applicable legislation on data processing.

Statistics Norway can grant access to the source data (de-identified or anonymised microdata) on which the statistics are based, for researchers and public authorities for the purposes of preparing statistical results and analyses. Access can be granted upon application and subject to conditions. Refer to the details about this at Access to data from Statistics Norway.

The objective of the Index of wholesale and retail trade is to describe the value and volume development in the wholesale and retail trade sector. The wholesale and retail trade sector consists of wholesale and retail trade and repair of motor vehicles and motorcycles, wholesale trade, except of motor vehicles and motorcycles and retail trade, except of motor vehicles and motorcycles. Wholesale is companies selling new and/or used goods to other companies, while retail trade is companies selling new and/or used goods to private households. Retail sales is executed from either a fixed or moveable sales outlet, a market place or via the Internet or mail order. Examples of retail sales are the sale of food, beverages, clothing, shoes, domestic electrical appliances, furniture, building equipment and so on.

The Index of wholesale and retail trade is the main component in the calculation of household consumption. The index have been published since 1936 and was until april 2021 named Index of retail sales. From april 2021 the index was broadened to also include sector 45 wholesale and retail sale and repair of motor vehicles and motorcycles and sector 46 wholesale, except of motor vehicles and motorcycles.

From November 1999, the sample of retail and wholesale companies has been supplemented gradually with turnover figures for chain stores reported directly by head offices. As a consequence, figures for newly established and closed down units for the above-mentioned stores are also more up to date.

Users include public and private sector agencies and organisations. Statistics Norway's national accounts statistics rely on timely production of the index of wholesale and retail trade. Other users include Statistics Norway's research department.

No external users have access to the statistics and analyses before they are published and accessible simultaneously for all users on ssb.no at 8 am. Prior to this, a minimum of three months' advance notice is given in the Statistics Release Calendar. This is one of Statistics Norway’s key principles for ensuring that all users are treated equally.

The index of household consumption of goods is published at the same time as the Index of wholesale and retail trade. The former is more extensive than the latter, which can lead to different developments in the two indices.

For calculating the volume indices of the Index of wholesale and retail trade the consumer price index is used for sector 45 wholesale and retail trade and repair of motor vehicles and motorcycles and sector 47 retail trade, except of motor vehicles and motorcycles while the price index of first-hand domestic sales is used for sector 46 wholesale, except of motor vehicles and motorcycles.

Changes in turnover are later compared with bimonthly wholesale and retail sales statistics which are based on the VAT register.

The statistics are developed, produced and disseminated pursuant to Act no. 32 of 21 June 2019 relating to official statistics and Statistics Norway (the Statistics Act).

The statistics are part of the national program for official statistics, main area Wholesale and retail trade and service activities, sub-area Wholesale and retail trade.

Regulation (EU) 2019/2152 of the European Parliament and of the Council of 27 November 2019 on European business statistics (EBS)

All establishments in the sector G, which consists of sector 45 wholesale and retail trade and repair of motor vehicles and motorcycles, sector 46 wholesale, except of motor vehicles and motorcycles (46.1 wholesale on a fee or contract basis is excluded) and sector 47 retail trade, except of motor vehicles and motorcycles (SIC2007: G.).

Monthly statistical surveys (simplified questionnaire), Statistics Norway's annual survey of wholesale and retail trade establishments (detailed questionnaire), the VAT register and Statistics Norway's Register of Establishments and Enterprises.

A sample of about 18 100 units is selected from the population of wholesale and retail trade establishments in the VAT register. This includes a sub-sample of 15 100 chain departments/stores with direct reporting from head office.

3 000 units is selected from other existing establishments, i.e. independent companies/shops and the remaining chain departments/stores not yet included in the sample above. This population is stratified according to size in terms of number of employees. The sample is adjusted as necessary to ensure a reasonably even geographical coverage and to incorporate available new information from the annual survey of wholesale and retail trade establishments and the bimonthly updated VAT register. In the latter case, the adjustments focus on diverging trends between turnover as measured by the sample and turnover as measured by a survey of the VAT register. In 1997, the sample was adjusted to take into account the variation of strata variances. The sample is rotated annually based on the second term of the turnover statistics. Establishments are retained in the sample for a maximum of four years unless they are part of a full coverage stratum.

Questionnaires are submitted electronically in the Altinn portal. If the respondents need help in filling in the questionnaire, Statistics Norway can be contacted by telephone. The establishments normally receive the questionnaire before the expiry of the survey month. The deadline is the 12th day of the following month. Failure to respond is subject to fines.

Editing is defined here as checking, examining and amending data.

Prior to the statistical compilation, arithmetic and logical checks are carried out. This procedure also includes comparing the results with other data sources, mainly the wholesale and retail trade statistics.

Results broken down by sector and stratum are compared with data for the previous period and the corresponding period in the previous year. If the discrepancy is considerable, the respondent is consulted.

A ratio estimator is applied to each stratum to inflate sample data to population level. The ratio estimator uses turnover figures from the VAT register as auxiliary variables.

The establishments are divided into identical units and newly established units. These concepts are defined in Notater 93/17 Detaljomsetningsindeksen (in Norwegian only).

Newly established units are enterprises that have been registered in the Central Register of Establishments and Enterprises after the last rotation of the sample. The turnover of these enterprises is based on estimates as this information is not available. The estimates are based on information about newly established units in the same period in the previous year. The average turnover per establishment is calculated for newly established units in the same period in the previous year. On this basis the turnover for all recent establishments for the period in question is aggregated. The estimates for newly established units are made on stratum and sector level.

Volume indices at four-digit sector level are calculated by deflating value indices directly by means of the consumer price index for sector 45 and 47 and the price index of first-hand domestic sales for sector 46. At two and three-digit sector level the volume indices are calculated as a weighted sum of the volume indices at the four-digit sector level using value shares in the reference year as weight.

The index is adjusted for seasonal variations applying the X12ARIMA method with non-fixed seasonal effects and multiplicative model. As a supplement to seasonal variations, the new model takes into account the effect of weekdays, fixed holidays as 1 st and 17 th May, Easter, Pentecost, Ascension Day and 1 st New Year's Day. 24th to 26th of December are considered as seasonal variations with a method developed by Statistics Norway.

See "About seasonal adjustment" at the end of "About the statistics".

Employees of Statistics Norway have a duty of confidentiality.

Statistics Norway does not publish figures if there is a risk of the respondent’s contribution being identified. This means that, as a general rule, figures are not published if fewer than three units form the basis of a cell in a table or if the contribution of one or two respondents constitutes a very large part of the cell total.

Statistics Norway can make exceptions to the general rule if deemed necessary to meet the requirements of the EEA agreement, if the respondent is a public authority, if the respondent has consented to this, or when the information disclosed is openly accessible to the public.

More information can be found on Statistics Norway’s website under Methods in official statistics, in the ‘Confidentiality’ section.

The Index of wholesale and retail trade has been published since 1936. In April 2021 the name changed from Index of retail sales, since the index was broadened to also include sector 45 wholesale and retail trade and repair of motor vehicles and motorcycles and sector 46 wholesale, except of motor vehicles and motorcycles.

A new version of Norwegian industry classification (SIC2007) has been implemented from January 2009. The toal index, as well as three-digit sector levels, has been recalculated back to 2000 according to SIC2007.

In January 2003, Statistics Norway altered the calculation method for the price index used to deflate the retail trade sector. At this time the price index was only used to deflate the retail trade sector 47, except of motor vehicles and motorcycles, but is from 2021 also used to deflate sector 45 wholesale and retail trade and repair of motor vehicles and motorcycles. The new method is based on the price development of product groups from the consumer price index, as well as on sector-wise product allocation. The product allocation comes from a survey with that particular focus. The current method used to calculate this price index is using the same methods used to calculate the deflator of the retail sales component in the index of household consumption of goods. Volume figures have been calculated back to August 1999 using the current deflator.

The volume indices at four-digit sector level are calculated by deflating value indices using the aid of the consumer price index for sector 45 and 47, and the price index of first-hand domestic sales for sector 46. From January 2002, volume indices at two and three-digit sector level are calculated as a weighted sum of the volume indices at four-digit sector level with value shares in the reference year as weight. In the past, volume indices at the two and three-digit sector level were calculated using immediate deflation of value indices.

In compliance with Eurostat regulations the reference year was changed to 2021=100 when indices for January 2024 were published. Prior to this the reference year of the retail sales index was 2015=100.

Measurement errors (the respondent supplies erroneous data) and processing errors (wrong interpretation of figures and letters during) may occur.

Three types of errors are common:

• The respondent does not supply turnover figures for the establishment, but for a part of the establishment or the enterprise that the establishment is part of.
• The respondent does not report data for the correct time period (calendar month).
• Figures are in the wrong unit of measurement (usually in NOK instead of NOK 1 000).

Reminders are sent to enterprises that fail to respond in time. Failure to respond is subject to fines. These enterprises are treated in the same way as enterprises that are not included in the sample. This means that the change in turnover applied to these enterprises is the average percentage change that is applied to their stratum. Non-respondents normally constitute about 3 per cent of the total sample at the time of publication.

The results are uncertain as they are based on information from a sample of enterprises. The sample is updated once a year. Sample errors may also occur as a result of errors in the information that the sample is stratified according to.

A revision is a planned change to figures that have already been published, for example when releasing final figures as a follow-up to published preliminary figures. See also Statistics Norway’s principles for revisions.

Revisions in previously published seasonally adjusted figures can take place when new observations (or revised previous observations) are included in the basis of calculation. The scope of the revision is usually greatest in the most relevant part (last 1–2 years) of seasonally adjusted time series. A corresponding revision in trends is also typical, particularly at the end of the time series. The extent of the revision of trends and seasonally adjusted figures is partly determined by the revision policy, see Section 4 of the European Statistical System (ESS) Guidelines on Seasonal Adjustment on the Eurostat website. For more information on the revision of seasonally adjusted figures, see the ‘About seasonal adjustment’ section in the relevant statistics.

Monthly and quarterly time series are often characterised by considerable seasonal variations, which might complicate their interpretation. Such time series are therefore subjected to a process of seasonal adjustment in order to remove the effects of these seasonal fluctuations. Once data have been adjusted for seasonal effects by X-12-ARIMA or some other seasonal adjustment tool, a clearer picture of the time series emerges.

For more information on seasonal adjustment: metadata on methods: seasonal adjustment

Due to our shopping habits The Index of wholesale and retail trade will vary from month to month. For instance the month of December shows higher sale than the rest of the months. This combined with the influence of how the Easter holiday varies between March and April and also the influence of movable public holidays make a comparison from one month to the next difficult. To adjust for these circumstances the Index of wholesale and retail trade is adjusted for seasonal variations, so the underlying development of the index can be analyzed.

Seasonally adjusted series

Both Value- and volume indices in the Index of wholesale and retail sales are seasonally adjusted and published down to the three-digit sector level. In total, this constitutes 48 seasonally adjusted series. These series include industries from all of NACE section G Wholesale and retail trade; repair of motor vehicles and motorcycles, including NACE division 45 Wholesale and retail trade and repair of motor vehicles and motorcycles, NACE division 46 Wholesale, except of motor vehicles and motorcycles and NACE division 47 Retail trade, except of motor vehicles and motorcycles.

Pre-treatment routines/schemes

Pre-treatment means adjusting raw data for calendar effects and outlier values before seasonal adjustment.

• A detailed pre-treatment of raw data is used. This means using models which are specially adapted for the pre-treatment of the raw data for a given series.

Calendar adjustment

Calendar adjustment involves adjusting for the effects of working days/trading days and for moving holidays. Working days/trading days adjustments are made for both the number of working days/trading days and for the varying composition of days from one month to another.

• Performing calendar adjustments on all series showing significant and plausible calendar effects within a statistically robust approach, such as regression or RegARIMA (a regression model with an ARIMA structure for the residuals). The regression variables for the calendar adjustment are adapted to reflect the working days, public holidays and so forth specific to Norway.

Methods for trading/working day adjustment

• RegARIMA correction – in this case, the effect of trading days is estimated in a RegArima framework. The effect of trading days can be estimated by using a correction for the length of the month or leap year, regressing the series on the number of working days, etc. In this case, the residuals will have an ARIMA structure.

Comments:

All series are adjusted for leap years with a separate variable.

For 3- and 4- digit industries of NACE Division G47 the calendar adjustment is modified such that only 1st of January, 1st of May and 17th of May are counted as extra Sundays. We do pre-treatment with a separate variable that captures all of the holidays related to easter. In addition, we also use a separate variable for Whit Monday. The method makes use of a standard 6-parameter model for each of the 6 shopping days: Monday to Saturday (TD).

For the total of NACE Division G47 the pre-treatment for the direct seasonal adjustment has been changed so that Maundy Thursday, Good Friday, Ascension Day, Whit Monday, Christmas Day and Boxing Day are counted as extra Sundays. For NACE Division G45 and G46, included the underlying 3-digit industries, Christmas Eve and New Year's Eve are also counted as extra Sundays due to marginal activity.

For the total of NACE Division G47 a 3-parameter model is used for Trading day adjustment, where trading days with similar shopping levels have been combined: Monday and Tuesday is one group, Wednesday to Friday another and Saturday is the final group.

For the other 2- and 3-digit industries of NACE Division G45 (both volume and value series) we use a 1-parameter model for trading day adjustment where all days, Monday to Saturday, are assumed to have similar shopping levels (here the number of Sundays is the contrast category). This 1-parameter model for trading day adjustment is also used for industry G46.7.

For the volume series of G45.2 a standard 1-parameter model for working day adjustment (WD) is used, where Monday to Friday is assumed to have the same level of shopping activity. Here the number of weekend days is the contrast category.

For the value series of G45.2 a 3-parameter model for working day adjustment is used, where working days (Monday - Friday) with a similar level of shopping activity are grouped together. Tuesday to Thursday make up one group, while Monday and Friday make up separate groups. The same 3-parameter model is used for the working day adjustment of NACE Division G46, as well as G46.3 and G46.9.

For G46.2, G46.4, G46.5 and G46.6 a standard 1-parameter model for working day adjustment (WD) is used. Here Monday to Friday is assumed to have the same level of shopping activity. The number of weekend days is the contrast category.

Correction for moving holidays

• For 3- and 4-digit industries of NACE Division G47 we adjust by estimating the duration of the effect of the moving holidays. This estimation is specially made for the Norwegian calendar. We use a separate variable for all of the holidays associated with Easter. This means whether the period from Maundy Thursday to Easter Monday entirely or partially is in March or April. In addition we use a separate variable for a Easter period of 7 days prior to Maundy Thursday, because an increased shopping level is observed in this period. Furthermore, a separate variable is used to estimate the effect of Whit Monday.
  
• For the total of NACE Division G47 we use a separate variable that captures whether Holy Saturday (the day between Good Friday and Easter Sunday) is in March or April. The Easter holidays are counted as extra Sundays, as is done with Whit Monday. To capture the increased level of shopping prior to Easter a pre-Easter variable spanning 6 days prior to Maundy Thursday is used.
  
• For NACE Division G45 and G46, including the underlying 3-digit industries, we do not use separate pre-treatment variable for Easter of Whit Monday since the moving holidays are counted as extra Sundays.

National and EU/euro area calendars

• A calendar based on Norwegian holidays is used.

Treatment of outliers

Extreme observations, also called outliers, are observations in the time-series deemed not to be normal.

• Outliers are identified automatically by the seasonal adjustment tool. The outliers are removed before seasonal adjustment is carried out, and the reintroduced into the seasonally adjusted data.

Comment:

The automatic search for outliers is only used to identify additive outliers, not level shifts. This is done to avoid unexpected jumps in the trend.

A constant significance limit is used to identify outliers automatically. The T-value is set to 4 as the threshold irrespective of the length of the period over which we are searching for outliers. This is recommended procedure by Statistics Norway.

For G45 and G45.1 there are multiple outliers that are likely related to changes in taxes and fees on the sale of cars. December 2020, 2021 and 2022 have specially high sale numbers and are defined as additive outliers, since we assume that this is not a lasting change in the seasonal pattern. November 2022 and January 2023 are also defined as additive outliers, so that these special observations shouldn't affect the estimation of the seasonal pattern.

Treatment of the COVID-19 pandemic 2020-2022

In the seasonal adjustment we have followed the recommended guidelines by Eurostat, which states that the effect of the COVID-19 pandemic shouldn't affect the estimation of the seasonal pattern. We have implemented the use of level-shift outliers for the entire period with national restrictions in Norway. This is the period from March 2020 until March 2022.

The exception from this are the short time-series for G45 and G45.1 where the treatment of the COVID-19 pandemic is done in a way that uses less parameters. For the period of March 2020 until March 2022 we have performed an automatic search after different types of outliers with a low critical threshold (t-value > 2). The different types of outliers are additive outliers (AO), level shifts (LS) and temporary changes (TC). For G45.1 the following outliers were identified and hardcoded (type and period): TC2020M03, LS2020M07, TC2021M04 and AO2021M05. For G45 was AO2020M08 also identified as an outlier.

Model selection

Pre-treatment requires choosing an ARIMA model, as well as deciding whether the data should be log-transformed or not.

• Automatic model selection by established routines in the seasonal adjustment tool.

Comment:

We use the procedure in the seasonal adjustment tool X-12 ARIMA called "pickmdl", which chooses the first and simplest OK ARIMA structure from a list of 5 models which produces a satisfactory result. If a satisfactory result is not achieved by these, then the alternative procedure "automdl" is used. This procedure uses the best ARIMA structure among many candidates. This strategy prioritizes model stability and low revision, especially with respect to yearly reviews, significantly more that the current default procedures in the seasonal adjustment tools X-12 ARIMA and JDemetra+. (The "pickmdl" procedure was previously the default procedure of X-12 ARIMA).

Decomposition scheme

The decomposition scheme specifies how the various components – basically trend-cycle, seasonal and irregular – combine to form the original series. The most frequently used decomposition schemes are the multiplicative, additive or log additive.

• Automatic choice of decomposition procedure.

Choice of seasonal adjustment approach

• Other

The R packages RJDemetra and pickmdl. More specifically the functions x13 and x13_spec from RJDemetra (https://cran.r-project.org/web/packages/RJDemetra/index.html) and the functions x13_pickmdl, x13_automdl and x13_both from pickmdl (https://github.com/statisticsnorway/pickmdl).

Consistency between raw and seasonally adjusted data

Do not apply any constraint.

Consistency between aggregate/definition of seasonally adjusted data

In some series, consistency between seasonally adjusted totals and the original series is imposed. For some series there is also a special relationship between the different series, e.g. GDP which equals production minus intermediate consumption.

Do not apply any constraint.

Direct versus indirect approach

Direct seasonal adjustment is performed if all time series, including aggregates, are seasonally adjusted on an individual basis. Indirect seasonal adjustment is performed if the seasonally adjusted estimate for a time series is derived by combining the estimates for two or more directly adjusted series.

• A mixed indirect approach is applied where seasonal adjustment is done by the use of different approaches and tools.

Comment:

And indirect approach is used for NACE Section G which is based on direct seasonal adjustment of NACE Division G45, G46 and G47. Direct seasonal adjustment is used on all remaining series. This means 2-, 3- and 4-digit level of G47 as well as 2- and 3-digit level of G45 and G46.

Horizon for estimating the model and the correction factors

When performing seasonal adjustment of a time series, it is possible to choose the period to be used in estimating the model and the correction factors. Correction factors are the factors used in the pre-treatment and seasonal adjustment of the series.

• The whole time series is used to estimate the model and the correction factors.

Comment:

This holds for all time-series except for G47. For G47 data from January 2012 and onwards is used to identify the best ARIMA and pre-treatment model, and estimating it. This is done so that the time-series shouldn't be too long, which can give sup-optimal choices and estimates towards the end of the series. This is in line with the guidelines from ESS regarding seasonal adjustment.

Starting with the publication of indices for November 2023 we changed the model from using data starting with January 2008 to use data starting with January 2012 for G47.

General revision policy

Seasonally adjusted data may change due to a revision of the unadjusted (raw) data or the addition of new data. Such changes are called revisions, and there are several ways to deal with the problem of revisions when publishing the seasonally adjusted statistics.

• Seasonally adjusted data is revised according to well-defined and publicly available revision routines and publishing calendars.

Comments:

Raw data is not revised.

Concurrent versus current adjustment

Partial concurrent adjustment: the model is identified yearly, while filters, outliers and regression parameters are estimated continuously as new or revised data become available. Outliers for the last year are re-identified continuously as new or revised raw data becomes available.

Comments:

Factors are estimated yearly.

Horizon for published revisions

• The period for revision of seasonally adjusted data is limited to 4 years before the latest revised raw data (or new data), while older seasonally adjusted data is frozen.

Evaluation of seasonally adjustment data

Evaluation of quality based only on graphical inspection and descriptive statistics.

Comments:

A model where the various quality indicators will be evaluated continuous/periodically in the future.

Quality measures for seasonal adjustment

For most of the series, a selected set of diagnostics and graphical facilities for bulk treatment of data is used.

Seasonal adjustment of short time series

• Short series treated by the seasonal adjustment procedure are dealt with in a special way.

Comment:

2- and 3-digit industries of G45 and G46 starts in 2015 with two years outliers years due to the COVID-19 pandemic. These time-series are as of now a bit short. This has lead to an alternative treatment of the COVID-19 pandemic for G45 and G45.1. Furthermore, a full 6-paramter model for trading-days adjustment was not considered for these industries.

Treatment of problematic series

• Problematic series are treated in a special way. Other series are treated according to regular routines.

Comment:

See the previous section.

Data availability

Raw data, pre-treated data and seasonally adjusted series are available.

All metadata information associated with an individual time series is available.

Press releases

In addition to raw data, at least one of the following series is released: Seasonally plus working day adjusted.

• Notat 2009/27 Dokumentasjon av sesongjustering i SSB
• ESS-Guidelines on seasonal adjustment
• US census: X-12-ARIMA-manual
• Nye US Census-baserte metoder for ukedageffekter for norske data
• Notat 2007/43 Ny metode for påskekorrigering for norske data
• Notat 2001/54 Sesongjustering av tidsserier, Spektralanalyse og filtrering
• Notat 2001/02 Innføring i tidsserier, Sesongjustering og X-12-ARIMA