The Statistical Business Register Israel study visit 28

The Statistical Business Register Israel study visit 28 -30 April 2014 Mr Steen Eiberg Jørgensen (sej@dst. dk), Deputy Head of Division, Business Register

Outline Our point of departure for registers The role of the SBR – most valuable player Data sources – the wider context Data model Usage of the SBR Organisation and cooperation patterns A fruitful approach Key learning points for Statistics Denmark 2

Our point of departure for registers 3

The SBR – most valuable player … • NSI management: Need for cost-efficiency – first you need to push – then users will start to pull • National Accounts: Need for coverage and coherence • Survey dept. : Need for “free” update of address books • IT dept. : Need for standardisation of systems • Methodology dept. : Need for a frame for extraction and optimisation of samples • Politicians: Need for reduction of administrative burden and productivity in the public sector 4

Main sources for the SBR The SBR is a satellite in a wider system of data users/providers Primary data providers Danish Business authority Central Customs and Tax Administration Secondary data providers Central ABR: the core Legal units Central Register of Persons Official Journal of the Danish State Statistics Denmark Danish Working Environment Authority Danish Labour Market Authority Local legal units Self registration via Internet 5

Basic data model for SD’s SBR Types of units comprised by Statistics Denmark’s SBR Statistical units Administrative units Enterprise group Enterprise Legal unit Establishment (LKAU) Locallegal unit (LKAU) VAT unit 1 VAT unit 2 VAT unit 3 Kind of activity unit (KAU) Responsibility: Statistics Denmark CABR / DCCA CCTA 6

Basicdatamodel for SD’s SBR - 2 Legal form ESA 2010 – sectorcode in SBR Activity (NACE) code Birth Death Statistical information • Turnover • Different employment data 7

Usage of SBR for surveys in Denmark SBR is used in 70 different statistics (structural, short-term, register based, sample based) in 8 divisions / 3 directorates: Either as the statistical business units, or by Adding characteristics to other types of statistical units Approx. 165 recurring extracts each year Approx. 50 pct. of NSI staff are on-line users of SBR and extract system Examples: • Expected investments in the manufacturing sector • Employment in construction sector • Stocks in manufacturing and whole sale • Purchases of goods and services in construction sector 8

Usage of SBR in Denmark – examples - 2 Sales of goods in the manufacturing sector Companies’ sales and purchases Consumer credits Inward FATS Financial sector companies Business demography Government finance statistics General accounts statistics (SBS) Statistics on utilities Structural employment statistics (obs) Retail trade index Tourism statistics 9

Usage of SBR in Denmark – examples -3 Forestry statistics Patents, design and innovation Business services statistics IT-expenses and IT-investments The public sectors usage of IT Private enterprises’ usage of IT Harvest of cereals International transport of goods National transport of goods Foreign trade in services Labour market stats, incl. wages and labour cost (obs) 10

SBR organisation – 2 • Cooperation between SBR and users: • Common house rules • SLAs (“what, why, how, who, when”) • Forum for problem-solving • Cooperation with suppliers: • IT (both current work and new projects): Prioritisation, methods, review, test etc. • Administrative sources • Statistical sources (incl. delegation) 11

Organisation – 3: SLAs WHAT WHY 1. Definition of extract Level of unit, cut-off, Compliance with variables, requirements fields/codes, format and guidelines 2. Delivery of population extract Extract according to Sample design specifications 3. Selection of sample 4. Update before sending out questionnaires HOW WHO WHEN Statistical divisions n - 1 week SBR team n Definition of method Ensure and criteria representativity Stratification, selection etc. Methodology n + 3 days division Maximize unit response Update for the SBR team units in the sample n + 3 days 5. Feedback of Update SBR corrections to SBR Optimize population Statistical division n + 14 days 6. Processing of updates Max. response Updates made SBR team Correct gross-up by SBR team n + 20 days 7. Prioritization and solving of problem cases Agreement on conflicting cases and “grey zones” Coherence between different statistics Define criteria SBR and statistician n + 20 days 8. Documentation of extract Ensure coherence and accessibility SBR team Initially and as needed 12

A fruitful mindset … 1. The SBR must adapt to surveys – and vice versa … 2. … so problems must be solved together 3. Implementing a SBR is a long process – priorities and milestones 4. Grasp the low-hanging fruits 5. Not everything can be automated – keep it simple 6. Not everything can (or needs to) be checked – trust the sources and prioritize error checking 7. Focus efforts on large and complex enterprises (“ 90/10” rule) 8. Clear out disagreements about data – and stick to the agreement 9. The register will never be perfect – just like statistics … 10. … but everyone gets more than they give 13