Description of Danish Practices in Retail Trade Statistics.

Transcription

1 Description of Danis Practices in Retail Trade Statistics. 1. Statistical units and reporting units Units and population Te enterprises in te Retail Trade Index are all legal units. In te Central Business Register of Statistics Denmark (CBR) te legal units are known by teir CVR-number. Tose units, wo ave teir main or secondary activity witin te sectors of retail trade, constitute te population. Te total number of enterprises in te sectors of te retail trade is estimated to be approximately 31,000. Tose 31,000 enterprises are delimitated to approximately 8,000 enterprises wit a turnover in excess of DKK 2,5 mil 1. From tis population a sample is made, wic is used as a frame for te calculations of te Retail Trade Index. 2. Data items and teir definitions Variables In te survey te enterprises are asked about teir sales, VAT included, to private ouseolds. Sale to private ouseolds is defined as retail trade. Statistical aims In te national publications, te index for eac sector of retail trade is stated as turnover index. For te tree main commodity groups, a deflated turnover index and seasonally adjusted index is also stated. Sectors of retail trade are grouped according to Danis Industrial Classification of All Economic Activities 2003 (DB03); te codes are sown in parenteses. A complete description of te sectors is given in DB03. Te sift from DB93 to DB03 from 1 January 2003 as caused some minor canges to a few sectors. Te delimitation of te population is based on te Central Business Register of Statistics Denmark and DB03. Te starting point is enterprises wit teir main activities witin retail trade. In addition, te enterprises wit teir main activity outside retail trade but wit teir second line of business witin retail trade are included. For te latter a recalculation is carried out, so tat only te part of te turnover connected to te retail trade is included. Furtermore, bakers are also included in te retail trade statistics, even toug teir sector in te trade classification belongs to te production industry. Te total turnover of te sectors included in te Retail Trade Index accounts for approximately 96 per cent of te total turnover in te wole retail trade sector. Some retail trade sectors are not included in te statistics, partly due to insignificant sales and partly due to some specific conditions in certain sectors. Since te survey is compulsory and te reminding procedure is intensive, te response rate is approximately 99 per cent, measured on turnover in final figures. 1 Tis reduction is made in order to curtail te administrative burden for te smaller enterprises

2 3. Data sources and data collection metods Te Retail Trade Index is based on teleponed or written response by te enterprises in te sample. Te Retail Trade Index began as a voluntary survey. Legal autority to collect te data is currently given by te Act on Statistics Denmark (Lov om Danmarks Statistik), Section 8. Te lifetime burden on te average respondent was for te year 2003 equivalent to 2.9 years of work. Coverage Te total number of enterprises in te population is estimated to be approximately From tis population a sample is selected tat is used as a basis for calculating te indices. Tis sample comprises units in te population, wic generate annual turnovers for at least DKK 2.5 million, VAT included. Te sample covers approximately 36 per cent of te enterprises and approximately 87 per cent of te total turnover from all te enterprises wit turnover of at least DKK 2.5 million in te sectors of retail trade. Sample/size class In te sample, all enterprises wit annual sales in excess of DKK 20 million, VAT included, are included. Te remaining enterprises are cosen optimally. Te population is divided into 4 size classes. Te size classes are: DKK million, 5-10 million, million and beyond 20 million. Among te enterprises wit less tan DDK 20 million in annual turnover, te size of te sample is determined by te turnover index. Te units wit retail trade as teir secondary line of business are cosen based on teir sare of te total turnover witin te sector of retail trade. Te limit of DDK 2.5 million in annual turnover for participating in te sample as been cosen to curtail te burden of te respondent for small enterprises. Te total selected sample consisted per 1 January 2005 of enterprises. Eac year te sample is renewed by approximately 1/3 of te enterprises wit an annual turnover between DDK million. Hence, it is assured tat te size of te sample is kept at te same level and te sample gives an accurate picture of te retail trade sector. Tis metod implies tat some enterprises can be exempted from te sample for some period. Collecting/measuring Te reported figures of te turnover are eiter typed in by telepone or obtained troug submission of completed questionnaires by te enterprises. Te questionnaires are posted at te end of te period for wic data are submitted. If te questionnaires ave not been returned witin a few days after te deadline, a reminder is sent, togeter wit a new deadline. If tis deadline is exceeded as well, te enterprise is reminded by telepone. Completion of te questionnaires is mandatory. However, some enterprises do not fulfil teir duty. In tose cases Statistics Denmark sends a registered letter, stressing out tat teir turnover must be reported witin a week. Oterwise, Statistics Denmark will notify te police. In te questionnaire, eac enterprise is asked about teir turnover, VAT included, to private ouseolds. Turnover to private ouseolds is defined as retail trade. Tis includes repairs undertaken at te premises of te enterprise; for instance, te repairs of watces are also included in - 2 -

3 te total turnover for te watc and clock stores. Supermarkets, discount stores and department stores sell a very wide variety of goods. In order to be able to estimate and publis te Turnover Index, wic is divided into tree main commodity groups: Food and Oter Basic Commodities, Cloting etc. and Oter Consumption Goods, te enterprises in tese sectors are requested to report te turnovers for tose tree main commodity groups separately. If te enterprise is unable to submit accurate figures, it is requested to estimate te sales stemming from eac group. Te submitted data undergo an error probability control. If te reported turnover deviates considerably from earlier records ten te enterprise in question is asked to investigate te correctness of te returned turnover figures. It is assumed tat not all errors in submitted forms are detected, and terefore te statistic is a subject to some uncertainty. Maintenance of te sample Due to te intensive reminder procedure, only a very limited number of enterprises in te sample do not return te completed questionnaires. Te sample is renewed eac year by refresing one tird of te sample. Hence, it is possible to update canges in te structure and in te development in te population s turnover. Furtermore, te annual renewal keeps te size of te sample at te same level and insures tat a level break will not occur in te enumerated turnover. Te survey is based on a sample and tus some uncertainty is attaced to te results. For te wole retail trade sector, owever, te uncertainty is considered to be small, since te coverage is extensive. A risk of a iger uncertainty is taking place in tose sectors, were te coverage is not always as extensive. Te montly estimation of development in turnover in eac sector of retail trade will terefore carry some uncertainty, especially in sectors tat mostly consist of enterprises wit an annual turnover of less tan DKK 2.5 mill. Uncertainty caused by wrongly reported figures and misunderstandings are sougt minimized by different means of probability control of te submitted figures. Basic material is kept and stored for approximately two years (in bot paper and electronic form). Good experience wit te data collection 5 different ways to report te turnover (fax, reporting by telepone registering service, reporting by calling directly by telepone, , te free post service), makes it easy to report. Difficulties wit data collection It takes time to sort out te reported figures reported by letters or fax. Delays or missing post is also a problem. 4. Indices and performance indicators Te Retail Trade statistics is publised bot as a turnover index and as a deflated turnover index. Te turnover index sows te development in te turnover in current prices. Tus it includes bot te development in volume and in price. Te turnover index is publised for all 49 sectors of te retail trade and for te 3 main commodity groups: Food and Oter Basic Commodities, Cloting etc. and Oter Consumption Goods. Te deflated turnover index, corrected for te development in prices, is a quantity index tat sows te development in te volume of sales. Te turnover index - 3 -

4 is deflated wit te weigted consumer price index. For more detailed information one can contact te person responsible for Prices and Consumption in te Department of Economic Statistics. Deflated turnover indices are not publised for all sectors of te retail trade, but only for te wole retail trade and te tree main commodity groups. Te indices for te tree main groups and te wole retail trade are also seasonally adjusted. Tis adjustment takes into account certain public olidays (e.g. Easter) as well as te number of trading days in eac mont. So far it is not possible to correct for all types of trading day effects. Calculation metod Te enumeration is based on te reported turnovers and te VAT-turnovers submitted to VAT Statistics. Information about VAT-turnovers is received every 3 mont from te Central Business Register of Statistics Denmark. VAT-turnovers are used bot to determine te optimal sample allocation and to improve te accuracy of te estimates. Te estimated Retail Trade Index is an expression for te development in te retail trade for enterprises wit an annual turnover in excess of DDK 2.5 mil. Tis development is also assumed to express te development in total retail trade. Te total estimated turnover is calculated as a sum of eac sectors estimated turnover. Te Retail Trade Index, based on te estimated turnover is calculated for every sector in retail trade, and for te total retail trade. Formulas for tese calculations can be seen in Statistical News, Service sector 2002:48. No absolute figures for te retail trade are publised. Anyone interested in te absolute turnover figures for eac sector or for te total retail trade, is referred to te statistics over business units registered for VAT settlement, publised wit industrial distribution. Te seasonal adjustment of te indices for te tree main commodity groups, te index for te total retail trade, and indices for te sectors tat are sent to EUROSTAT, is performed by te use of X-12 ARIMA, wit te user surface Demetra. Te seasonal adjusted series are corrected in suc a way, tat te annual sum of te seasonally adjusted series equals te annual sum of te original series. If detailed information on te seasonal adjustment is needed, one can contact te person responsible for tis statistics or te Metodological Unit of Statistics Denmark. Te figures in te publication News from Statistics Denmark are based on a ig percentage of reported turnovers; ence, tere are only minor differences between te provisional and te final figures tat are publised in te Statistical News articles. Good experiences witin te calculation Tis year te software package CLAN (te collection of SAS macros) is used to estimate te turnovers. CLAN being a very flexible SAS tool, estimates population totals and wat is most important, it calculates te corresponding standard error of te estimates. Terefore, it as become easier to judge and improve te quality of te estimates. Apart from te results tat are written to te existing Excel files, te SAS program also produces different output tables on te estimates in te SAS output window. Tose tables are very relevant in order to control te quality of te index. It makes it easy for instance to ceck weter all te respondents actually ave responded. At te very end of te program, tere is an additional program tat can be run in order to view canges in te responded turnover witin te selected sector from te previous mont. Tus all tables in te SAS output windows are successfully used for analyses of te estimates witin te selected sectors, as tey contain te current results of all te necessary calculation and data treatment. In addition, in te Excel output tables, te direct estimates, te ratio - 4 -

5 estimates and te estimates based on canges from te previous mont, are sown togeter wit teir coefficient of variation for eac sector. Te difficult issues witin te calculation One of te major issues witin te calculation is tat te Central Business Register of Statistics Denmark updates only every 3 d mont. It implies tat we receive te information about te VAT turnover wit an average delay of 3 monts. Suc a delay in te update of te auxiliary variable can cause an overestimation for instance in cases of mergers among larger enterprises in te sectors of te retail trade. 5. Distributive trade statistics and compilation of te national accounts Oter business trend statistics also sow te development in te retail trade. Below is given a brief description of te relationsip between te Retail Trade Index and a number of oter statistics, wic can be used to assess te retail trade development. VAT statistics Besides being an all-inclusive statistic, te VAT statistic is different from te Retail Trade Index statistic in te sale and unit concepts tat are used. In te VAT statistic, te total turnover is te total domestic turnover for te enterprises tat ave retail trade as teir main activity. Wile te Retail Trade Index only covers te turnover to private individuals. In te Retail Trade Index te enterprises wit te retail trade as teir secondary activity are also included. Te VAT statistics are based on current payments of VAT. Tis means tat canges in population ave in principle an immediate effect on te figures in te VAT statistic. Te same is not te case wit te Retail Trade Index. Consumption of goods in te National Accounts Statistics Te Quarterly National Accounts state private consumption including te ouseolds consumption of certain types of products, for wic development in turnover are included in te Retail Trade Index. In te National Accounts Statistics, figures are stated regarding te quarterly development in consumption by product type. Oter statistics Apart from te above-mentioned statistics, wic can be used to assess te development in te total retail trade, a number of oter statistics partly cover ouseold consumption: For te purpose of business trend valuations, montly statistics are made for sales of newly registered motor veicles, based on information from te Central Register of Motor Veicles. Annual statistics based on samples are made for sales in a number of service trades (for example retail trade in personal computers, etc., wolesale sales of office macines, etc.) A number of associations witin te retail trade prepare statistics on developments in sales based on data from teir own members

6 Forms of dissemination Current publications: News from Statistics Denmark, Service sector (Statistical News) and Main Indicators. Yearbooks: Statistical Yearbook. Statistics Denmark s StatBank Denmark at

7 ANNEX. Statistics Denmark Some additional issues regarding service assignments in statistic production, Retail Trade Index Generally about statistics: Montly enumeration. 49 sectors of retail trade. Enterprises wit turnover in access of DKK 2, 5 million, VAT included. Service assignments Description Resources (work ours per year) 1 Selection of te sample Selection of te sample is made by Te Department of Wo selects te sample Metods 2 Preannouncements - Does te enterprises receive announcements about participation in te sample in advance No enterprises are contacted in advance. Oterwise, we are following te policies of Statistics Denmark regarding te enterprises tat participate in more tan five different statistics. 3 Information to enterprises - Wo prepares te information Wo is te information sent to If everyone receives te same information Wat kind of information is sent Tere are 2 kinds of questionnaires, wic are sent to different sectors. 90 % of respondents ave to report only one figure for teir turnover. Te remaining 10 % must split teir turnover into 3 main commodity groups: food, cloting and oter consumption goods. Every year wit te renewal of te sample, te questionnaire, te information folder, te introduction letter, te advertising paper tat tells about te opportunities to use te telepone for reporting te turnover and an example of te statistic we publis are sent to respondents. Te contest of te introduction letter depends on weter te enterprise is new in te sample or is already familiar wit te procedure. Te following mont only te questionnaire, te information folder and te result of statistic are sent to te respondent. Te questionnaire, te free envelope and an example of te publised statistics are sent every last working day of te reference period (a mont)

8 4 Procedure of sending te questionnaires Wo prints te questionnaires Wo puts te questionnaires into te envelopes Wo fulfils te questionnaires wit te key data (addresses, journal number, etc.) Questionnaires are fulfilled wit te key data automatically. Te data comes from our own respondent-register (Oracledatabase). Packing and sending te envelopes takes place externally (Handicappers Enveloping Bureau). Printing te questionnaires: Data collection and registration How can te data be collected Wen is te collected data registered App. 25 % of te respondents respond by using te telepone reporting system. Oters use eiter fax, mail, post or call directly to us. Registration of te data, manual setting in and searcing for te inaccuracy occurs simultaneously. Oracle is used. Post receiving and sorting: 300 Manual setting te figures in: 1000 How is te collected data registered How is te data registered Wat IT-systems are used Weter te questionnaires being sort before putting in arcive 6 Scanning Are te questionnaires made appropriable for Questionnaires are made appropriable for scanning. As soon as te scanning program is adjusted, te reported turnover will be scanned. scanning 7 Answering te questions from te respondents Typical questions Te typical questions are: wy as te enterprise to report and weter te frequency of reporting can be reduced. Te persons wo collect all te data are also te contact persons wo answer all te questions. Answering questions: 50 te Wo answers te It is possible to acieve te respite. questions Can te enterprise get a permanent respite - 8 -

9 8 Reminder procedure How many enterprises ave to be reminded Wat is te reminder procedure Wo is reminded and after wat criteria 1. Reminder: pr. letter (te deadline is 1 week) tereafter, large enterprises are reminded troug te telepone before te calculation of te index. Only if te respondent as not answered witin te 3 mont, e is reminded wit a registered letter. Afterwards te police are noticed. Tere are app. 300 registered letters a year. (After te 1.quarter app. 281 registered letters are usually sent, mostly because te new respondents don t know tat tis reporting is statutory) Police cases How many police cases a year 5-10 police cases a year Micro searc for inaccuracy (te individual respondents) Te absolute /probable inaccuracy Electronic control Manuel control How te inaccuracy is corrected (respondent contact) Tere is an automatic control wen using te telepone reporting system. Te figures tat are not proved automatically ave to be cecked manually (it is usually about 30 % of te telepone response). If tere is found an inaccuracy ten te respondent is contacted by te telepone. Wen te inaccuracy is corrected, te explanation note about te correction is made. Wen te data is registered, bot electronic and manual control is performed. Control and correction of te inaccuracy: 650 Te sare of inaccuracy Control of all te respondents 11 Macro ceck (control of te results) Control after te enumeration 12 Reporting back to enterprises Statistics is send back 13 Register update Update Te enumeration is made twice a mont. After 30 days te provisional figures are publised and after 50 days te final figures are publised. Tere is made a visual control of te calculated indices. Enterprises receive no receipt. Statistics is send to all enterprises. Te respondent register updates by canges in key data. Information about te sectors-, address- and cvr- cange is reported furter to te 14. Department (National Accounts) and to te Central Business Register of Statistics Denmark (CBR). Sectors canges are reported to CBR. Current update of CBR:

10 Te Estimation Procedure 1. Estimation of te turnover Every mont te total retail trade turnover Y is estimated. Generally, it can be estimated wit te direct estimate: direct t N t direct t Yˆ ( ) ( ) ( ) = y = Y ˆ i, m were te population size in stratum 2 is given as N. Te number of respondents in stratum is given as m n, were n is te number of sampled elements in stratum. Te value of m can cange during te year due to nonresponse. Te estimate of Y can be improved by using auxiliary information. In our case te information of te VAT-turnover (wic is received from te Central Business Register of Statistics Denmark (CBR)) can be used advantageously to improve te accuracy of te estimates of te target variable. Wit only one numerical auxiliary variable, te ratio estimator is te obvious coice of estimator in te estimation process. Te ratio estimator is a special case of te regression estimator. Te use of tis auxiliary information is justified wen te auxiliary variable correlates wit te variable of interest te turnover y. Tis is indeed te case in te retail trade. Denoting te auxiliary variable as X, te ratio estimate of te turnover is defined as: were and i s pop( t) X ratio t g t Yˆ ( ) ( ) = Y ˆ ( t) g, g Xˆ Yˆ Xˆ ( t) g ( t) g = = N n g i s g N n i s g are te direct estimates of te turnover and te VAT-turnover, respectively, in sector g of te retail trade at time period t. Te expression pop( t) X g = x is te population total of te VAT-turnover in sector g at time period t. As we receive te information from te CBR every 3 mont, and tus only ave information about te quarterly VATturnover for eac enterprise in te sectors of te retail trade, te variable x t i is te average annual VAT-turnover for eac enterprise and is calculated from te four latest available 3-monts periods. i U g y ( t) i x ( t) i ( t) i 2 Te detailed description of te definition of te stratum is given below in te section Stratified sample

11 1.2. Estimation of te index Te Retail Trade Index at time period t is defined by te turnover at time t wit te turnover in some basis period. At present, te basis period is te year 2000; te turnover in te basis period is defined as te simple average of te 12 montly turnovers in te basis year. Te retail trade index is tus estimated by ˆ ratio( t) ˆ Y It = 100, Y0 were Y 0 is te average turnover in te basis period. Te above expression for te index is in te literature referred to as a direct index. Alternatively, te index at time t can be written as a cain index, wose estimate is given by ratio( t) ˆ ˆ ˆ Y I t = It 1, ratio( t 1) Yˆ tat is, by te (estimated) index in te previous period multiplied by te cange between te estimated turnovers in two mont t 1 = t 1 and t 2 = t: Yˆ Y ratio( t2 ) 1,2 I ˆ =. ratio( t1 ) Especially wen dealing wit a rotating panel design, te latter way of estimating te index is very useful. At present, te cain index is used. 2. Stratification Tere are 49 sectors in te retail trade, te indices for wic are estimated every mont for a national publication. Furtermore, indices of 14 additional sectors are calculated on bealf of EUROSTAT. Te strata in te sample are determined as follows: Sector of retail trade x size class (te level of annual VAT-turnover 3 ), were size classes are: DKK million, 5-10 million, million and more tan 20 million (see Table 1 for an overview over size classes). Te VAT-turnover for te previous point in time 4 determines te class size of eac enterprise k. Te average 5 VATturnover for eac enterprise serves furtermore as auxiliary information in te estimation process and is denoted by te variable named x. Eac stratum is a combination of bot te size class and te sector, so as tere are (49+14) sectors and four size classes, tere are in principle (49+14)*4=252 in all. In reality, tere are defined 2 extra strata. Tose two extra strata include enterprises wit retail trade as teir secondary activity (Benterprises) and te kind of enterprises tat ave a network of stores and report te turnover for all tose stores as an aggregate single figure (F-enterprises). It is important to single tose two types of enterprises out in separate strata, because tey are treated differently in te estimation process. For eac enterprise k U te corresponding stratum is known. Table 1. Overview of size classes. Size class Total VAT-turnover according to register (x i) 1 Annual sales btw. DKK millions 2 Annual sales btw. DKK 5-10 millions 3 Annual sales btw. DKK millions 4 Annual sales in excess of DKK 20 millions 3 VAT-turnover (or total sales) is te turnover tat enterprises report to VAT Statistics. Te turnover tat is reported to Retail Trade statistics is called for turnover. 4 Usually te last year 5 Annual turnover divided by 12 to estimate te average turnover per mont

12 3. Sampling procedure Te sample S is a rotating panel sample. Te total sample size is about 3000 enterprises. Every January 1/3 of te enterprises is refresed. Every January, bot enterprises in te old sample, as well as te newly included enterprises are observed. In te period of February to December, only te new sample is observed (tose wo remained in te sample and te newly selected enterprises). Enterprises in size class 4 (wit an annual VAT-turnover of DKK 20 million or more) are all included in te sample. Tey are not replaced, as tey all participate every year. Te same is true for te enterprises wit te retail trade as a secondary activity (B-enterprises) and te enterprises tat report as a network (F-enterprises). Te B-enterprises are included in te sampling frame U if te sare of teir total VAT-turnover witin te retail trade is above te cut-off of DKK 2.5 million. All B-enterprises in te frame U participate in te sample. Te sample is drawn using te stratified simple random sampling (STSI) wit an optimal design, were te optimal allocation is based on te observations from te last year prior to te new sampling year. 3.1 Optimal allocation (Neyman allocation) Te key ingredient in te optimal allocation is te product N S, were N is te population size in stratum, and S is te stratum standard deviation of te variable Z on wic is optimized. As we estimate te index of te retail trade turnover, te optimization of te sample sould be performed wit respect to te index I. However, te index is not linear: Te index I cannot be found as te sum of te estimated strata indices, tat is, Iˆ Iˆ and terefore Neyman s formulas cannot be used directly. In addition, since te cain index is used in te estimation process, we wis to optimize on canges in te index from one mont to te next instead of on its absolute value. Moreover, wen estimating mont-to-mont canges, te gain of using auxiliary information and te ratio estimator in te estimation process is small, compared to te direct estimator. Terefore, te sample is optimized on te quantity Tis is equivalent to optimizing on te variable I ˆ 2,1direct Yˆ Yˆ direct( t2 ) direct( t1 ). e = y I y, 2,1 ( t2 ) 2,1 ( t2 ) i i i as explained in Houbiers and Fangel (2004). Overall, te quantity S in Neyman s formula is given by were 2 1 2,1 2,1 ( S ) = ( ei e ) n 1 i S 2,

13 e = n 2,1 1 is te stratum average of e 2,1, wic generally is non equal to zero. As te new sample as to be drawn only once a year, te average value of te standard deviation over te 12 monts (11 mont to mont canges) in te past year is calculated: 2 S i S e 2,1 i 11 1 ( t, t 1) = 11 S t = 1 S. Tus, te optimisation strategy for tis sample is te following: Te strata corresponding to size class 4, te B- and te F-enterprises: opt n = N Te strata corresponding to size classes 1,2,3: opt = ~ NS n min N,max n,3, NS were n ~ is adjusted suc tat te total sample size is equal to n. It follows from te formulas tat te sample size n in stratum is larger wen stratum contains more elements N, and/or wen S is larger, tat is, wen te variability of te residuals in stratum is larger. In practice, for strata containing 3 or less enterprises, te wole stratum is observed

14 List Over Sectors of te Retail Trade. Classifications. In te national publication, te following classifications of sectors of retail trade are used: Grocer s sops ( ) All-nigts sops ( ) Supermarkets ( ) Discount stores ( ) Retail sale of fruit and vegetables ( ) Retail sale of meat and meat products ( ) Retail sale of fis, game, crustaceans and molluscs ( ) Bakers sops and retail sale of bread, cakes and flour confectionery ( and ) Retail sale of cocolate and sugar confectionery ( ) Retail sale of alcoolic and oter beverages ( ) Retail sale of tobacco products ( ) Retail sale of ceese ( ) Retail sale of oter specialised stores wit food ( and ) Varity and department stores ( and ) Dispensing cemists ( ) Perfumery sops ( ) Cemists (not licensed to dispense medicine) ( ) Retail sale of textiles ( ) Retail sale of ladies cloting ( ) Retail sale of men s cloting ( ) Retail sale of men s and ladies cloting ( ) Retail sale of baby articles and cildren s cloting ( ) Retail sale of footwear ( ) Retail sale of leater goods ( ) Retail sale of furniture ( ) Retail sale of furnising fabrics ( ) Retail sale of kitcen utensils, glass and cina ( ) Retail sale of electric ouseold appliances ( ) Retail sale of radio- and television goods ( ) Retail sale of records, CDs, cassettes, etc. ( ) Retail sale of musical instruments ( ) Retail sale of ardware ( ) Retail sale of building materials ( ) Retail sale of paints and wallpaper ( ) Retail sale of books, newspapers and stationery ( ) Retail sale of carpets ( ) Retail sale of watces and clocks ( ) Retail sale of watces, clocks and jewellery ( ) Retail sale of jewellery ( ) Retail sale of glasses ( ) Retail sale of potograpic equipment ( ) Gifts sops ( ) Retail sale of sports goods ( ) Retail sale of toys and games ( ) Retail sale of bicycles and mopeds ( ) Retail sale of computers, standard software and office macinery ( and ) Retail sale of telecommunications equipment ( )

15 Florist s sops and retail sale of plants and seeds ( and ) Retail sale via mail order ouses ( ) Groupings are made by use of DB03; te codes are sown in parenteses. A complete description of te sectors exists in Danis Industrial Classification of All Economic Activities 2003 (DB03), Statistics Denmark Te sift from DB93 to DB03 from te 1 of January 2003 as caused minor canges to a few series. For te national publication te following grouping of main commodities are used: Food and Oter Basic Commodities cover te total turnover at grocers and sops wit specialised sales witin food and oter basic commodities, inclusive department of food in supermarkets, discount stores, and department stores. Cloting etc. consists of te turnover in stores wit specialised sales witin cloting, footwear and textiles inclusive supermarkets, discount stores, and department stores sales of tese goods. Oter Consumption Goods cover te turnover mainly for stores wit sales of equipment for ome and leisure togeter wit te supermarkets, discount stores, and department stores sales of tese goods. In Eurostat publications, anoter grouping of te sectors are used and te following sectors are included: , , , , , , , , , , , , ,