Adaptation of the SBR to changing external conditions

Adaptation of the SBR to changing external conditions Søren Schiønning Andersen, Statistics Denmark (ssa@dst. dk)

Outline • Background and purpose • The three basic registers - our point of departure • The integration of the SBR with the CABR • The sources • Recent challenges • Pros and cons • Key differences between SBRs and ABRs • Overall learning points

Purpose To describe: • The interface and integration between our SBR and the main sources, especially the CABR • Recent challenges to our SBR and how we have tried to cope with these challenges • How these changes have affected our ability to fulfil the objectives of the SBR (coverage, quality, usage, costs) • Our main learning points from these challenges

Our point of departure re registers

Basic data model for our SBR

Main sources for the SBR

Six recent challenges to our SBR 1) Change to a key administrative source 2) Change to the underlying business model 3) Adaptation of the SBR to 1) and 2) 4) Challenges to content from political strong users 5) New supra-national requirements 6) In-house requirements for improved productivity

Challenge 1: Change to a key source Cause: • e-Income from CCTA register replaces old source • Obligatory monthly employment data at LKAU level Effect: • Re-design of process and IT system (1 man-year) • Improvements to timeliness/frequency, relevance and accuracy • Increased usage of SBR Conclusion: • Pro-active communication at an early stage is key • Data definitions must comply with statistical needs • The CCTA must have self-interest in data quality • Follow CCTA’s project all the way to implementation

Challenge 2: Change to the business model Cause: • “Web-reg” due to new administrative usage of CABR • Cost reductions in the CABR • Change in the underlying perception re BRs Effect: • LLUs were no longer followed over time – fundamental change to the business logic • High quality potential, but also high risks. We will see … • So far, it has meant more work for SD Conclusion: • Keep statistical concepts and needs on the agenda – try to give as much as possible in return • Exploit the advantages and avoid the disadvantages

Challenge 3: Adaptation of the SBR Cause: • Necessary response to challenge 1 and 2 Effect: • Functionality re LKAUs had to be built-up in SBR instead • Less tight relation with CABR – more freedom … • Work processes changed from administrative to statistical units • Extremely costly compared to available resources Conclusion: • Keep it (more) simple – it is very hard to get resources to change very complex systems that only benefits SD • I. e. we need to be more realistic and modest when we define requirements

Challenge 4: Strong political pressure on ABR Cause: • Pressure to re-use data in order to reduce burden • Digitalisation and productivity of the public sector • Data must fulfil more purposes – also from strong players Effect: • Data will have direct effects on data subjects • The relative weight of statistical needs will diminish • New incentives and sources of errors are introduced • Net quality effects are difficult to assess Conclusion: • Monitoring of new initiatives • Proactive communication and advice to admin. users • The overall system must remain sustainable

Challenge 5: New supra-national requirements Cause: • New EU Regulation with additional requirements (EGs) • Data needs are not covered by the CABR • No other administrative source (share holder register) Effect: • We must rely on commercial data • Data on EGs and MNEs are already in high demand Conclusion: • Ensure that new supra-national requirements for the SBR are incorporated in a coming administrative register • Avoid parallel systems becoming permanent

Challenge 6: How to do more with less … Cause: • Recurring cost reductions in the NSI • Currently, our SBR do not cover all units in agricultural surveys - a separate farm register is maintained Effect: • Integration of missing agricultural units into the SBR • Actuality and accuracy of data on farms will increase • Coherence will improve Conclusion: • The main sources/systems must be exploited to maximum extent • Redundant data and systems should be discontinued • Traditions and “cultures” are difficult to change

Pros and cons … summing up Benefits / advantages Costs / disadvantages High coverage (depending on requirements for legal registration) High quality (if data are validated by the administrative body and updates are well coordinated) Reduces costs Reduces administrative burden on business Supports frequent statistics (if updated on a current basis) Potentially more timely data (depending on technical set-up and administrative processes) Administrative definitions can deviate from statistical needs and definitions Problems regarding match and consistency Use of different classifications (and different use of the same classifications) Limited possibilities for collecting supplementary data (because of burden considerations and division of responsibility/competence) Reluctance among enterprises regarding data exchange (necessitates assurance of confidentiality and “documentation” of positive cost/ benefit ratio) Dependency on providers Vulnerable to political and/or administrative changes Potentially less timely data

Key differences between SBRs and ABRs Aspect SBR ABR Purpose Supports production and dissemination of aggregate statistics Supports binding legal or administrative decisions about individuals Content Focus on economic phenomena – at institutional and productive level Focus on legality and responsibility of individual operations Complex approach to units (More) simple approach to units Consequences Neutral - no direct consequences Defines rights and obligations – direct consequences Relation to policies Indirect Direct Political weight “Light” – intangible benefits “Heavy” – tangible implications The SBR is a “bi-product” The ABR is the “main” product Time perspective Basis for coherence and comparability over time – complex approach Basis for decisions with effect from date of registration – simple approach Quality philosophy Multi-dimensional approach More simple approach Data are accurate when equal to a real value – which is often unknown Data are accurate when equal to the self-perception of the unit “Sufficient quality“ is unclear “Sufficient quality“ is (more) clear

Key learning points for SD In order to better fulfil our role and objectives we must: • Manage our partnerships better – we are not strong enough alone • Communicate proactively – we cannot wait for others to contact us • Always be part of the solution – not part of the problem • Manage our risks better – otherwise they seem to manage us • Keep things simple – balance ambitions with abilities

Thank you for your attention! Any questions or comments?