Deliverable 4.3. Riga Workshop: ESSnet on Consistency. Project on statistical units. - Report -

Transcription

1 EUROPEAN COMMISSION - EUROSTAT ESSnet on consistency of concepts and methods of business-related statistics 2010 project on statistical units n Deliverable 4.3 Riga Workshop: ESSnet on Consistency Project on statistical units - Report - Authors: Sarmite Prole, Svetlana Jesilevska Final Version 1.2

2 Content Introduction... 3 Welcome and presentation of ESSnet on Consistency workshop... 4 ESSnet on consistency... 4 Session 1: What do we have? and What do we need? st part: What do we have?... 4 Introduction and presentation of ESSnet on consistency Project on statistical units... 4 Results of the qualitative part of the Inquiry on identification of inconsistencies between Member States... 6 Enterprise... 6 Kind-of-activity unit Local unit Local kind-of-activity unit National Accounts and Regional Accounts Results of the quantitative part of the Inquiry on identification of inconsistencies between Member States Outline on identification and evaluation of inconsistencies in regulations for business related statistics Results of the vignettes on the test of a new definition of enterprises nd part: What do we need? Improved, new, operational, etc. definitions of the main statistical units for European business statistics Session 2: Where do we go? Which statistical units lead to Business Statistics best fit for economic analysis? Inconsistency problems from the point of view of National Accounts Identify «enterprises»: the profiling procedure Statistical units: Avoiding mistakes, identifying a consistent approach. From the statistical concept of autonomy to the statistical definition of enterprise Round up and conclusion Annex I Workshop Agenda Annex II - List of participants WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

3 Introduction Character of ESSnet Consistency The ESSnet on consistency of concepts and methods of business and trade-related statistics is divided into three activities: Statistical units (WP1) Target population, frames, reference period, classifications and their applications (WP2) Characteristics and definitions (WP3) The work on statistical units is the first task of the whole ESSnet. All activities refer to the consistency item. Objectives and organization of the workshop ESSnet on Consistency - Project on statistical units workshop (workshop) is carried out within the framework of the ESSnet on consistency of concepts and methods of business-related statistics (WP project on statistical units). The workshop was successfully held on the 19 th and 20 th of June, 2012 at Hotel Avalon in Riga, Latvia. The goal of the workshop was to present the work done at the ESSnet project on consistency of Concepts and Methods of Business and Trade-related Statistics and to open discussion on the needs for revised statistical unit s definitions. Project on statistical units comprises the following five National Statistics Institutes (NSIs) of Austria (STAT), France (INSEE), Ireland (CSO), Italy (ISTAT) and Latvia (CSB). The co-ordinator of the WP1 is ISTAT. The workshop hosted by the Central Statistical Bureau (CSB) of Latvia with intensive support from the countries involved in WP1 and Eurostat. The workshop was divided in two sessions (Annex I): Session 1 was divided in two parts. The content of the 1 st part was covered under question What do we have? where was given overview of the results of the Inquiry on statistical units. There were presentations on the results of the Inquiry on identification of inconsistencies between Member States on Enterprise and Legal unit; Kind-of-activity unit, Local unit, Local kind-of-activity unit and on National Accounts/Regional Accounts. An outline on identification and evaluation of inconsistencies in regulations for business-related statistics was also presented. The identification and evaluation of inconsistencies in regulations is carried out within the ESSnet on Consistency Project on statistical units. The last presentation outlined the results of case studies on enterprise definitions on the test of a new definition of enterprise. The 2 nd part was covered under Question What do we need? presenting improved, new, operational, etc. definitions of the main statistical units for European business statistics. Session 2 Where do we go? touched upon opinions of NSIs on which statistical units lead to Business Statistics best fit for economic and inconsistency problems from the point of view of National Accounts. The final presentations covered enterprise profiling procedure and definition of enterprise for statistical purposes. WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

4 Experts from Business statistics and National accounts from each Member State were invited to the workshop. Around 58 experts from 21 European countries and Eurostat joined the workshop. (See Annex II -List of participants). Kaspars Misans, vice president of the CSB of Latvia, introduced the Agenda and emphasized the problem of consistency on statistical units where the statistical outputs of in different domains cannot be compared because the definitions, concepts and methodologies are partially or even completely different. Welcome and presentation of ESSnet on Consistency workshop Aija Zigure, president of the CSB of Latvia, welcomed everyone to the ESSnet on Consistency workshop. She underlined that it had been a pleasure for the CSB to host this workshop and also emphasized the meaning of ESSnet as it is a way to develop new projects, where Member States interested in a specific domain can actively collaborate on common tasks and then disseminate the results. She mentioned that the entire ESSnet on consistency operating within the MEETS Programme and ESSnet on consistency of concepts and methods of business-related statistics project on statistical units fell under objective 2 of the MEETS programme aiming at the achievement of a streamlined framework for business related statistics. ESSnet on consistency (Hans-Eduard Hauser, EUROSTAT) Hans-Eduard Hauser from Eurostat in his presentation showed the necessity to harmonise concepts and definitions across all business-related statistics and Member States. He also outlined expectations of Eurostat from the workshop on statistical units and ESSnet on consistency in general. Session 1: What do we have? and What do we need? Chair: Sarmite Prole (CSB, Latvia) Sarmite Prole outlined the content of the1 st Session. She briefly explained that the session would be split in two parts. The content of the 1 st part was covered under question What do we have? and the 2 nd part was covered under question What do we need? 1 st part: What do we have? Introduction and presentation of ESSnet on consistency Project on statistical units (Giuseppe Garofalo - Coordinator of the project on statistical units, ISTAT, Italy) The presentation of Giuseppe Garofalo, Coordinator of the project on statistical units, covered the following: Identify and evaluate inconsistency regarding: o Inconsistency in statistical units applied in different statistical domains of the system of business and trade-related statistics according to the European legislation WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

5 o Inconsistency in the application of the statistical units definition in different Member States o Inconsistency of the existing statistical units for the development of new European statistical domains Make proposals for adjustment in EU regulations He also explained the general aim of the questionnaire: To identify and analyze the following typologies of inconsistencies: o Inconsistency in statistical units applied in different statistical domains of the system of business and trade-related statistics according to the European legislation o Inconsistency in the application of the statistical units definition in different Member States To take qualitative and quantitative measures The inconsistencies were identified and evaluated in sections through qualitative and quantitative measures for the following units: Legal units/ Enterprises Local units Kind-of-activity units (KAU) Local kind-of-activity units (LKAU) and for the following domains: Business register (BR) Structural Business Statistics (SBS) Short-term Statistics (STS) National/Regional Accounts (NA/RA) The structure of Inquiry on the Statistical units is described in Figure 1. Figure 1: The structure of Inquiry on the Statistical Units He mentioned that the response rate in the qualitative part of inquiry was very high (about 90% for all sections 1-6 regarding BR, SBS and STS domains) and for NA/RA the response rate was about 71%, shown in Table 1. WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

6 Table 1: Response rate by Sections The compilation of inquiry was sent to 27 Member states of the European Union and to the EFTA Countries. Response was received from 29 countries. Results of the qualitative part of the Inquiry on identification of inconsistencies between Member States There were presented results of the Inquiry on identification of inconsistencies between MS on Enterprise; Kind-of-activity unit, Local unit, Local kind-of-activity unit and on National Accounts/Regional Accounts. Enterprise (Monica Consalvi - ISTAT, Italy) Monica Consalvi explained in her presentation the main results of the qualitative part of the inquiry on the statistical unit enterprise. Regarding results of inquiry she briefly described what kinds of administrative/ statistical sources were used in the respondent countries for the implementation of the enterprises in BR and for the detection of the main characteristics of this unit. She emphasised that the majority of the questions mainly referred to the definition of the enterprise as stated in Council Regulation (EEC) No 696/93 of March 1993 on the statistical units for the observation and analysis of the production system in the Community (Regulation 696/93 on the statistical units). The definition has been divided into three main parts and the application of each one has been investigated. WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

7 Enterprise is strictly dependent on how the legal unit is identified in view of the fact that the consistency of the BR with the definitions of statistical units in the regulation can depend on national legislation, especially in economic and fiscal area. For Legal units, more than 61% of countries use more than one source and the average number is 2.7 per country. As regards enterprises, all countries use more than one source even to detect the characteristics of the unit. Exactly as expected, the fiscal sources are the most used for the variable turnover (71.4%) in the same way as the Social Security is for employment (75%). It is worthy of note that a considerable number of countries make use of surveys as a primary source to detect the values of the main characteristics and this is true above all to assess the economic activity code (42.9%), shown in Table 2. Table 2: Respondent Countries by main administrative/statistical sources used to implement the enterprises in the BR and main detected characteristics (% out of 28 respondent countries) Within the presentation also mentioned that the application of each part of the definition has been researched using specific questions. Table 3: Respondent Countries by methods to apply "Part 1" of the Enterprise definition As regards Part 1 of the definition, most of the countries perform some analyses to apply the concept of productivity; about 42% of countries say they use more than one WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

8 method to find out that a unit is an organizational unit producing goods or services. 13% of the countries consider each identified legal unit as a productive unit without further analysis, while more than 80% of countries say they perform further analyses to understand how the unit is actually productive, above all by using sources that could give this information (45%) or by using thresholds of certain variables (42%) or profiling methods (35%). Other methods are also relevant (19%) in this multiple choice. The most used variables for threshold are employment (61.5%), turnover (61.5%) and other sources like income and wages (53.8%). Table 4: Respondent Countries by features of autonomy to apply "Part 2" of the Enterprise definition As regards Part 2 of the definition, more than 61% of countries say they consider more than one issue among the practical concepts stated on Eurostat BR Recommendation Manual 2010, in the Chapter 19D (Operational rules for the enterprise Defining the enterprise), with an average number of 1.7 issues per country. It was assumed that the concept of autonomy in decision-making could be connected to the following issues: for about 68% of the countries it is relevant that the enterprise has accounts at its disposal and 42% consider if the operations of the enterprise are managed in an integrated manner. Not so many used the third concept, as only 39% of the countries consider the market orientation of an enterprise as an important issue. It should be noticed that 26% countries consider also some other issue as fundamental in the concept of autonomy. As regards Part 3 of the definition, half of the countries declare to follow only one approach and half of the countries follow precisely two of them. Figure 2: Relationships between Legal Units and Enterprises 1 Legal Unit Enterprise % (14) Legal Unit m Legal Unit. Legal Unit Enterprise % (12) Legal Unit Enterprise m Legal Unit. Legal Unit Enterprise. Enterprise n 3.2 % (1) Other relationships 6.5 % (2) WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

9 In particular, 45.2% countries has a one-to-one relationship between legal units and enterprises, thus, unexpectedly, more than half countries perform some activities to combine together legal units for the identification of the enterprise. The majority combine more than one legal unit in the same enterprise (38.7%), while only one country consider it possible to split one legal unit in more than two enterprises. 6.5% declare to use some other methods to identify the enterprise as the smallest combination of legal units, for example in some cases the ancillary units that serve another unit are identified but the accounts are not consolidated. To understand better which are the operations performed in case of combination/split of legal units and how far they are applied, some questions were asked to the subset of 15 countries where the link between Legal unit and Enterprise differs from a one-to-one relationship, explained in Table 5. Table 5: Coverage of application of Part 3 of the Enterprise definition With regards to this subset of countries, there is a different coverage in the application of the operative criteria to combine/split Legal Units: only five countries broaden them to the whole population, while the other ones restrict the analysis to a relevant subset of units in the BR, having a large dimension (20%) or a particular legal form (30%), a specific economic activity (10%), or limiting the analysis just to the units involved in surveys or under manual investigations (40%). Figure 3: Typologies of Legal units (LeUs) to be combined/ split to apply Part 3 of the definition WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

10 As regards the typologies of Legal units (Figure 3) to which the rules to obtain an enterprise combining legal units could be applied, the discussed situations are the ones stated in Eurostat BR Recommendation Manual, Chapter 19E (Operational rules for the enterprise Applying the definition). The most common situation is that the legal units carrying out vertically or horizontally integrated activities are combined to form a single enterprise (73.3%) but there are also 60% of the countries that take into consideration the legal units set up to hold the assets of two or more enterprises within an enterprise group. Taking into account the information shown in the last three tables, together with the additional description of the activities performed by the Member States, it was possible to classify better the countries in terms of their treatment of legal units. According to the high/low coverage of the treated units in the BR and the number of typologies taken into consideration, about 52% of the countries perform some activities to combine legal units in the same enterprises when they considered them not autonomous, but actually only 30% of the countries carry out a complete analysis and apply in practice the procedure of the profiling approach in a strict sense. They represent 20% of the total enterprises in Europe and 18% of their employment. Table 6: Treatment of Legal units (LeUs) Summary of the results of the qualitative part of the inquiry on statistical unit enterprise : The use of threshold (turnover and/or employees) is the main criterion to identify enterprise from legal units The most often used criterion for autonomy is an enterprise has accounts at its disposal (68% of the NSIs) Profiling methods seem to be used by 11 (35.5 %) countries Combination of LUs (using profiling or other methods) are performed by 15 countries Regarding the inquiry results: 8 countries applied strictly for the combination of LUs: 19.8 % in terms of total European enterprise, 18.0% in terms of total European employment Kind-of-activity unit (work was done by the CSO, Ireland) Giuseppe Garofalo explained results made by the CSO. Highlighting the main points, he explained the KAU definition according to Regulation 696/93 on the statistical units: The kind-of-activity unit groups all the parts of an enterprise contributing to the performance of an activity at class level (four digits) of NACE Rev.1 (2) and WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

11 corresponds to one or more operational subdivisions of the enterprise. The enterprise s information system must be capable of indicating or calculating for each KAU at least the value of production, intermediate consumption, manpower costs, the operating surplus and employment and gross fixed capital formation. He described the role of KAU in the spine of Business Regulations. He also provided information on the countries which have KAU in each statistical domain. In the Business Register (not required by Regulation) 11 countries implement this unit, 17 countries use KAU to produce SBS statistics (required by regulation), 13 countries use KAU to produce STS Industry statistics (required by regulation) and 2 countries use KAU to produce other STS statistics (not required by regulation), shown in Table 7. Table 7: Respondent Countries by presence of KAU in the statistical domains According to results of the inquiry 29 countries answered the questions on KAU (Iceland and Liechtenstein did not); 15 of these 29 countries do not have KAU or Local KAU on their Business Register. The presenter mentioned that 17 countries can observe/estimate some or all variables for KAU in SBS, using information available in the BR or from other sources; 3 countries definitely cannot observe/estimate any variables for KAU in SBS; 9 countries Do not identify KAU. Figure 4: Which countries have KAU and/or Local KAU on Business Register? Summary of the results of the quantitative part of the inquiry on statistical unit KAU : The majority of the NSIs are not compliant with the KAU definition of Regulation 696/93 on the statistical units that is mandatory for SBS and STS WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

12 Only 8 NSIs (corresponding to 25.8% of the European Enterprises and to the 33.4% of the European Employment) observe/estimate the compulsory 6 variables Limiting the KAU definition to the observation of value of production and employment the coverage increased to 61.9% and 62.4% Local unit (Pierrette Schuhl, Sylvie Mabile - INSEE, France) Sylvie Mabile offered analysis of the Local units concerning the questions of the qualitative part of the inquiry. She explained a local unit definition according to Regulation 696/93 on the statistical units where local unit is defined as an enterprise or part thereof (e. g. a workshop, factory, warehouse, office, mine or depot) situated in a geographically identified place. At or from this place economic activity is carried out for which save for certain exceptions one or more persons work (even if only parttime) for one and the same enterprise. One can find questions about Local Units mainly in Module 3 in different sections of the inquiry. Module 5 was also reviewed, to look at relationships between LUs and LKAUs. Some information was also issued from filter questions in Short Term Statistics questionnaires, mainly in section C or D dealing with Turnover and Production in Industry. But the results of the second module on Regional Accounts will not be developed here, as only 3 countries answered Mainly local units to the question What is the «statistical unit» for which appropriate basic data are available to compile Regional Accounts in your country? (Single answer) Not any further indication was provided with these answers. I. The Business Register experts point of view (Module 3 section A) In order to identify inconsistencies of the applied definition of Local Unit in different MSs and different statistical domains, the same third module was included in five sections of the web survey to NSIs (to be filled in by BR and STS experts, but not included in SBS section B, as Local Unit is not a statistical unit to be used in SBS regulation: see KAU or Enterprise) I.1 Response Rates Table 8 reports the global response rate of modules 3 in those five questionnaires. The information on Local units is mainly issued from BR experts (27 countries, section A). Only one country has submitted a module on Local Units in section D (STS in industry, Production) (see part III for further details on STS). Table 8: Response rate to Module 3 (Local Units) in all sections (number of countries) Number of modules 3 SECTION submitted A (BR) C (STS) D (STS) E (STS) F (STS) No Yes 27* 1 (empty) Total STS sections: C=Turnover (industry); D=Production (industry); E=Turnover (Retail trade & repair); F=Turnover (Other services) WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

13 The response rate is very high, above 90%, except for the requested description of other geographical details in the open question 3.4, where the response rate is only 50%. The feedback area in the Local Unit module was filled-in by more than 32% of countries (10 countries). Table 9: Response rate to Module 3 (Local Units) - in Section A (BR) Number of Respondent Qualitative questions in Module 3 - Local Units respondent rate (%) countries 3.1. List the main administrative/statistical sources used to implement the local units in the BR: Name of the source Fill in the administrative definition of the identified unit: Reg. 696/93 states that "at or from" the place where the local unit is situated "economic activity is carried out for which one or more persons work for one and the same enterprise". How does NSI identify local units? Methods of further analysis (multiple answer) Description of method used Do you have local units without employment? List the main sectors/typologies for which you have local units without employment in your BR: Which is the maximum geographical detail used to identify a local unit of an enterprise? Description of "Different geographical details" or "Other" Is it possible in your NSI to georeference the local units? Do you store more than one economic activity of the local units in your BR? If 3.6=Yes, list the main administrative/statistical sources that permit to store more than one economic activity of the local unit: If 3.6=Yes, does NSI identify these economic activities for the whole population of local units in the BR or just for a subset of them? Description of the subset Fill in your Local Unit "operative" definition: Please provide us your feedback: I.2 Results To implement the local units in the BR, 66% of countries use more than one source. Most countries use one or two sources, the average are 2.2 sources per country. Only one country indicates no source. Table 10: (Question 3.1) List the main administrative/statistical sources used to implement the local units in the BR (number of sources listed) Number of sources Number of countries Percentage Total Number of sources 0 1 3% % % % % % 15 Total % 64 Mean = 2,2 The sources used may be administrative or statistical sources: 4 countries use only statistical sources; 11 countries, only administrative sources and 13 countries use the two types of sources ( mixed sources ). WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

14 The main administrative sources used are Social security sources (11 countries), Legal sources (10 countries) and Tax sources (7 countries). Statistical sources include Structural Business Surveys, Profiling sources and others business surveys. But some sources are not well defined. Table 11: (Question 3.2) Regulation 696/93 on statistical units states that at or from the place where the local unit is situated economic activity is carried out for which one or more persons work for one and the same enterprise. How does NSI identify local units? No of countries % No Response Each place is considered a local unit without further analysis Further analysis (multiple answer): By using sources that identify exactly a local unit with employment By using thresholds of one or more variables (e.g. persons employed, employees,...) By using more complex methods (deterministic/probabilistic models) Other Each place is considered a Local Unit without further analysis by 13 NSIs (42%), but the majority of countries (16 countries, 52%) use further analysis. The following question supported multiple answers, but 87% of these experts stated only one kind of analysis, mainly answering by using only sources that identify exactly a local unit with employment (11 countries). Besides, 5 countries use thresholds and/or more complex methods. Question 3.3: Do you have local units without employment? This question investigates the existence of a minimum amount of employment in each unit. The answers suggest that there is, in practice, some confusion between the employment concept and the salaries (or wage earners) one. Regulation 696/93 on statistical units mentions for which one or more persons work : this criterion is probably not applied in the same way by all countries, depending on the available information. In fact, a majority of countries (19) do have LUs without employment in their BRs. But for 3 of them, there is probably confusion between employment and salaries, and for 6 of them, the answer underlines problems in BR quality or the absence of information about local employment. In practice, the employment criterion seems not to be a necessary condition to define a legal unit in the Business register. Question 3.4: Which is the maximum geographical detail used to identify a local unit of an enterprise? Question 3.5: Is it possible in your NSI to georeference the local units? Almost all countries (25) use addresses to identify the LUs. But one mentions the registering of (X, Y) coordinates, one monitors Zip code, one monitors Municipality and one does not register any geographical detail. A majority of countries is able to georeference the LUs (17 countries). Two more countries mention that they could be able to do it in the short term. Question 3.6: Do you store more than one economic activity of the local units in your BR? If Yes: WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

15 Question 3.7: List the main administrative/statistical sources that permit to store more than one economic activity of the local unit Question 3.8: Does NSI identifies these economic activities for the whole population of local units in the BR or just for a subset of them? A majority of countries (18) store more than one economic activity, but only 8 of them store it for all LUs and 10 of them store it only for a subset of their LUs. The subset is defined by surveys, sector (agriculture) or If known by legal source. Conversely, one country out of three (11 countries) store only one economic activity per LU in its BR. Sources used to store more than one activity of LUs in BR are: only statistical sources, for 3 countries only administrative sources, for 5 countries and mixed sources, for 10 countries Question 3.9: Fill in your Local Unit operative definition Most countries (15) give Eurostat s definition or similar as their operative definition: 8 countries use administrative definitions 3 countries use LKAU 2 countries use active locations and 1 country uses LU=Enterprise Some of the countries that use administrative definitions may also use the results of survey(s) to define Local Units. II. Relations between LUs and LKAUs (BR section- modules 3+5) As regards the relation between Local Unit and Local KAU, some results could be obtained looking at the answers in Module 3 (concerning Local Unit) and Module 5 (concerning KAU). The following results are not directly issued from an explicit question in the web questionnaire, but from the confrontation of the different answers and more generally, from the qualitative answers in Module 1 (BR section). Figure 5: Relations between LUs and LKAUs For most countries (16), the questionnaire answers suggest that LUs are not explicitly different from Local KAUs. Moreover, some countries do not implement different statistical units for Local Units and for Legal units (or Enterprises). But in other WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

16 countries, the definition of LU is clearly derived from the LKAU, or the reverse: LUs are defined by aggregation of LKAUs in 5 countries, whereas LUs are surveyed to define LKAUs in 3 countries. One country implements more complex methods to define LUs and 4 countries did not give any information about LKAU. III. Short Term Statistics (STS) sections In the beginning of sections C, D, E, F 1 of the survey sent to short term statistics experts, one and the same filter question was asked: How does your NSI identify the population of statistical units observed in your Short- Term Statistics? (Single answer) 1. Enterprise 2. Local Unit 3. KAU 4. LKAU For each item, 4 choices were prompted, and further modules were eventually required, depending on the answers of the following: Local Units are the ones identified in the Business Register without any further manipulation Local Units are identified by SBS Local Units are identified by other sources Units are identified using both the BR and other sources Whatever unit, if the first choice is checked (without any further manipulation), no further module is asked. Notice that none of these domains have the Local Unit as a mandatory Unit: in STS, the mandatory unit is KAU in Industry, and the Enterprise in other activities. That explains why the information on local units is mainly issued from BR experts: no country was prompted to answer any module 3 in SBS sections. Table 12: Response rates of the filter questions in STS sections (and Local Unit answers); ( How does your NSI identify the population In Short Term Statistics? ) STS Section Identification of Statistical Units N of countries % observed in STS C Industry (turnover) No response Local Unit Total (responses) D Industry (production) No response Local Unit Total (responses) E Retail trade.(turnover) No response Local Unit 0 0 Total (responses) F Other services (turnover) No response Local Unit 0 0 Total (responses) Sections C and D relates to Industry (Turnover and Production), section E to Retail trade and repair (Turnover) and section F to Other services (Turnover). WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

17 Only 2 to 4 countries (over 31) did not answer the filter questions of the Short Term Statistics questions, depending on the section. But when they did, only 4 chose the Local Unit item in the industry sector (13% of countries, 14% of responses), and none in retail trade or other services. In services, the mandatory unit is the enterprise. So, it is not surprising that no country declares using the local unit in these activity sectors (sections E and F). Nevertheless, some countries do use LKAUs (4 countries in retail trade and repair, and 2 of them in other services ). In manufacturing, the mandatory unit is the Kind of Activity Unit, but this unit is not mandatory in the statistical business register. Some NSIs may use proxies for KAUs, as local units. 4 countries report LUs for STS in industry, and they are the same in the Turnover (C) and the Production (D) sections. Note that in one country, the LKAU statistical unit was replaced in 2008 by the Local Unit, for administrative simplification reason. In the Turnover section, all 4 countries using Local Units for statistics in Industry didn t mention any further manipulation to define local units. But in the Production section, one country reported extra sources as SBS and Prodcom. Figure 6: Statistical units observed for STS in Industry (Production) and in Retail Trade (Turnover) Conclusion: Local Units are essentially evaluated by BR experts Local unit is not relevant for statistical units for STS, except some cases for industrial production, and if yes, information comes directly from BR In practice: LUs always in BR, georeference is often available and possible for a large majority of countries, but they are not always distinct from LKAUs (and in some cases, from Enterprises) Multiple administrative or statistical sources are used, but employment is sometimes missing at the local level, so it is not a compulsory criterion to define local units 2/3 countries monitor more than one activity (but usually not for all units), and the scope of the register may vary, including or not non market activities The relevance with other domains within this inquiry was not tested (labour market, environment). WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

18 Local kind-of-activity unit (Inga Oksentjuka, Liga Pranca CSB, Latvia) Inga Oksentjuka started her presentation with reference to definition of Local kind-ofactivity unit stated in Regulation 696/93 on the statistical units: Local Kind-of- activity unit is a part of kind-of-activity unit, which corresponds to local unit. Each KAU must have at least one Local KAU which corresponds to a local unit. Each enterprise should have at least one Local KAU. She provided insight into inquiry questions on statistical units. As regards the BR section of the questionnaire, 29 countries answered the questions on Local KAU, even though this unit is not mandatory in BR Regulation. 19 of them do not have Local KAU on their Business Register and in accordance with the information from the Inquiry, 10 countries have implemented Local KAU in the BR. As regards the negative answers, when the NSIs were asked about the main reasons not to implement the Local KAU in their BR, the most frequently mentioned reason was in general the lack of information. Actually, out of 19 respondent countries who have not implemented Local KAU in their BR, five of them mentioned the lack of information to cover the population as the main reason, which is 26% of the total number of countries. Another 21% of countries affirm not to have enough information, due to the lack of input sources. Another important reason is the lack of importance given to this unit, both from a legal point of view and in terms of users needs. Four countries (21%) mentioned as a reason the fact that identification of the LKAU is not mandatory in Regulation (EC) No 177/2008 of the European Parliament and of the Council of 20 February 2008 establishing a common framework for business registers for statistical purposes (Business register Regulation 177/2008) and thus it is not required in other provisions of law (either on the national or European level). Two countries (11%) have no internal and external users for Local KAUs and have no request for them. 5% of respondent countries mentioned that Local KAU was not a priority and another 5% said there are no strong needs for implementing Local KAUs in the BR. Results shown in Table 13. Table 13: Implementation of the Local KAU in the BR The presenter mentioned that in general the administrative information and the statistical sources have quite the same importance: they are used respectively in 90% WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

19 and 80% of the countries. In one country the units themselves inform the NSI about new Local KAU. Figure 7: Administrative/statistical sources of the Local KAU (1) Regarding Figure 8, the 40% of countries who implemented Local KAU use only information from Statistics as the main source, which includes Statistical surveys like Employment survey, SBS survey, BR survey and information from other statistics. 20% of the countries use only administrative sources, while the majority (40%) use both administrative and statistical sources at the same time. Figure 8: Administrative/statistical sources of the Local KAU (2) Figure 9 shows that 20% of the respondent countries use KAUs and local units for the identification of Local KAUs. The main source for this approach is the statistical surveys. The 10% of respondent countries mention that they identify Local KAUs directly from KAUs, using statistical surveys as the main source but also making use of administrative sources like the Tax register. Another 10% identifies Local KAUs from Local units, above all using branch approach. In most of the countries (60%) the identification of Local KAUs starts using other sources like surveys, different registers or legal units. WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

20 Figure 9: Starting point of Local KAU identification 70% of the NSIs are capable of observing/estimating the variable employment and 40% observe and/or estimate the variable value of production. Another 40% use other variables and in most cases it is the turnover. Only one country has some information available on all the six variables, while at least 60% of the countries are not able to observe/estimate any other variable. Figure 10: Available variables to identify the Local KAU The countries who implement the Local KAU in the BR were also asked if the identification involves just a subset of units or if it takes into account the whole population of local units in the BR. As shown in Figure 11, 40% of respondent countries identify Local KAU for all local units, while 60% succeeds in act upon a subset of them. Figure 11: Coverage of the implementation of the Local KAU The main criteria to choose the subset is by using the variable persons employed or performing some operations on certain specific NACE classes that are considered relevant by the NSIs. WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

21 The countries that have the Local KAU implemented in the BR were also asked to provide an operational definition of this unit, with the aim to obtain a precise description of the operational criteria used to define and identify this unit. According to the answers to this question, 10 countries were classified into three groups. Figure 12: Operational definition of Local KAU Summary of the results: 19 countries do not have LKAU, because this unit is not mandatory in BR. 10 countries have implemented LKAU in BR, out of them only 1 country has all 6 variables for KAU: o Value of Production o Intermediate consumption o Manpower costs o The operating surplus o Employment o Gross fixed capital information In STS and SBS the use of LKAU is not mandatory. The discussions on results of the qualitative part of the Inquiry on identification of inconsistencies between Member States in BR, STS and SBS: It was noted to draw attention to cases where one entrepreneur is involved in several enterprises. Discussions: Clarification regarding sole proprietors: According to Regulation 696/93 on statistical units legal units include natural persons who are engaged in an economic activity in their own right, so one can conclude that sole proprietorship is legal unit Regarding with inquiry results the majority of countries sole proprietors considered as enterprise. Only the NSI of the Netherlands uses the method of splitting the legal units for identifying of the enterprise WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

22 National Accounts and Regional Accounts (Kerstin Gruber - STAT, Austria) Kerstin Gruber explained the main results of the National Accounts (NA) and Regional Accounts (RA) questionnaire. She mentioned that the focus laid on the collection of the following information: Statistical units used in NA and RA The most important data sources regarding statistical units Quantitative effects of statistical units used in NA and STS and SBS the use of LKAU is not mandatory Regarding results of inquiry: Core data on NA/RA base regarding KAU/LKAU units in the SBS Implementation of the statistical unit KAU/LKAU differs widely from country to country Status quo of the implementation needs some further effort in NA/RA; It has clearly turned out that detailed information below the enterprise level is essential to produce harmonized and comparable macroeconomic data sets No further discussion. Results of the quantitative part of the Inquiry on identification of inconsistencies between Member States (Danila Filipponi ISTAT, Italy) Danila Filipponi presented the results on the quantitative part of the Inquiry on identification of inconsistencies between Member States. The aim in collecting quantitative information on the statistical units was to measure the inconsistency if any of the statistical units among Member States where the considered statistical units were Enterprises and KAU. Concerning Enterprises, the idea was to evaluate the effect of the actions carried out by each country for the identification of the statistical unit as defined by the Regulation 696/93 on the statistical units. The impact of the applied treatments has been quantified considering, for each country, the dissimilarity between the distribution by NACE code of the legal units and the distribution of the enterprises. The measure used to assess the dissimilarity between the two distributions is the Index of Dissimilarity. It is a measure that vary between 0 and 1, where the higher is the number, the more unlike the two distributions are. The formula for computing the Index is: D = 1 2 K i= 1 s f i f r i d Where k is the number of modality in the distributions and f i is the frequency of the modality i in the distribution d, with i = 1K,, k and d = r, s. The index D has been evaluated for three different dimensions: number, employed and turnover. Figure 13 shows the index by country and type of treatment. WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

23 Figure 13: Index of dissimilarity (of units, employed and turnover) by type of treatment and country Figure 14 provides a means of summarising the data (Countries, class of dissimilarity for the units and treatments) in two-dimensional graphical form. Figure 14: Correspondence analysis for Index of dissimilarity of units, type of treatment and country In order to explicate the index of dissimilarity of the distributions by NACE before and after the treatment as a function of the applied treatment of units and in relation with WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

24 the countries, it has been performed a Correspondence analysis. treatments have been divided as: The applied Label in graphs 1. Part1 NO Part3 NO 2. Part1 YES Part3 NO 3 Part1 YES Part2 YES 4 Part1 NO Part2 YES Description of treatment No treatment of LeUs Thresholds or other methods to identify the productive units Both combination of LeUs and methods to identify the productive units Only combination of legal units The conclusions for the enterprises are the following: The application of the PART 1 of the definition of Enterprise (identification of productive activity threshold-) brings: o big variation between legal units and Enterprises in terms of unit o small variation in terms of employment and turnover The application of the PART 3 of the definition of Enterprises ( combination of units ) brings variations in terms of employment and turnover The distribution of Legal units with respect to the distribution of Enterprises (in Units): o decrease the incidence of NACE L and S (<-1) o high increase of G, M and F (>+1) o medium increase of C, H and Q (+0.5 to +1) o all the others are stable variation Concerning KAU, the idea was to compare the distribution by NACE of the enterprise and KAU, considering as measure number, employed and turnover. Figure 15 shows the Index of dissimilarity of Number, Employed and Turnover by country and definition of KAU. Figure 15: Index of dissimilarity of Number, Employed and Turnover by country Table 16 show that out of the 29 countries that filled in the questionnaire, only 12 countries filled in the KAU quantitative section. Using a statistical model (GLMM) and WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

25 the information on secondary activity, it has been estimate the number of KAU for the missing country. Table 14 shows the number of KAU and enterprises in EU/EFTA by NACE. Table 14: Number of KAU and enterprises in EU/EFTA by NACE The conclusions for the KAU are the following: The results of the KAU analysis it is partial because of missing data for most of the biggest countries There are not evident differences between the Enterprise NACE distribution and KAU NACE distribution, even if the analysis carried out at section level could be partial Discussions: The delegates requested some clarification on the presentation, among statistical unit - enterprise and KAU, the availability of segment information and mentioned the national needs for collecting information. Relationships between legal unit and enterprise were also discussed. It was emphasised that the use of combined legal units is of small importance to figures on employment and turnover but concentration of the production is higher. The delegates expressed opinions on the medium-size enterprises and their role in common statistics, especially on determination of complex enterprises. In the field of the interpretation of KAU in the countries it was decided that it must be determined in which cases KAU is used and in which cases - the secondary activity. Regarding data from Euro Groups Register only two countries (the United Kingdom and the Netherlands) have a significant number with more than one legal unit per enterprise. WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June

26 Outline on identification and evaluation of inconsistencies in regulations for business related statistics (Svetlana Jesilevska - CSB, Latvia) Svetlana Jesilevska presented an outline on identification and evaluation of inconsistencies in regulations for business related statistics. She mentioned that the identification and evaluation of inconsistencies in regulations for business related statistics fell under objective of WP1 Statistical units of the Project ESSnet on consistency of concepts and methods of business-related statistics. The presentation was focused to explain what kind of statistical units are used in different statistical domains and their compliance with Regulation 696/93 on the statistical units. Svetlana Jesilevska presented four types of statistical units: Units as defined in Regulation 696/93 on the statistical units, with clear reference to Regulation 696/93 on the statistical units (SBS; STS; Structure of Earnings Survey/Labour Cost Survey; Prodcom; Tourism; Waste Statistics; Vocational Training Statistics; Statistics on Information and Communication technology; Foreign-Affiliates Statistics; Environmental expenditure; Units according to Regulation 696/93 on the statistical units, but with exceptions and rather weak link to Regulation 696/93 on the statistical units (Research and Development; Innovation; Energy Statistics) Well defined specific units, without reference to Regulation 696/93 on the statistical units or with reference in a very indirect way (External Trade Intrastat; Extrastat; Balance of Payments; Foreign Direct Investment Statistics; NA) Undefined units (Job Vacancy Statistics (the compilation of statistics shall be based on business units, but no definition is given and any reference to Regulation 696/93 on the statistical units is also missing); Labour Cost Index (Regulation 696/93 on the statistical units is quoted but there are no provisions which of the units have to be used); Labour Force Survey (the individual person is the central statistical unit, regardless of whether the sampling unit is an individual or a household. Information usually has to be collected for all individuals of the household). Figure 16: Percentage of statistical domains according to type of statistical unit Only half of statistical domains according to requirements stated in regulations and manuals are based on statistical units as defined in Regulation 696/93 on the statistical WP1: ESSnet on consistency Project on statistical units, Riga, Latvia, June