Register-based Statistics

Anders Wallgren and Britt Wallgren worked for many years in teaching and research as senior lecturers in statistics at Stockholm, Linköping and Örebro universities. They have also worked as consultants for several Swedish manufacturing industries. In connection with these consulting assignments, they began working with corporate administrative registers and developed statistical methods for business intelligence. After their academic career, they came to Statistics Sweden and worked for about 20 years at the Department of Research and Development. They worked mainly with survey methodology for the statistical registers which form the basis for the production of official statistics in Sweden. Since 2011 they have provided courses and consulting services on behalf of the Inter-American Development Bank in several Latin American countries and in the Caribbean.

• Index to Charts xiPreface xixChapter 1 Censuses, Sample Surveys and Register Surveys 11.1 The national statistical system 21.2 The traditional census-based system 31.3 New sources: Administrative registers and Big data 51.4 Basic concepts and terms 71.4.1 What is a register? 71.4.2 Databases, records and observations 81.4.3 What is a register survey? 101.4.4 A register survey: The Income and Taxation Register 121.5 New demands and opportunities require new methods 141.5.1 A new paradigm is necessary 141.5.2 New statistical methods 151.5.3 The basic principles of register-based statistics 181.6 Preconditions for register-based statistics 191.6.1 Reliable administrative systems 201.6.2 Legal base and public approval 211.6.3 Political support to strengthen the statistical system 23Chapter 2 The Transition to a Register-based Production System 252.1 First obstacle: How to gain access to microdata? 262.2 Protection of privacy and confidentiality 262.3 Second obstacle: How to take care of dirty data? 292.4 The new production process 302.4.1 Contacts with administrative authorities 312.4.2 Metadata have a new role 312.4.3 Anonymisation of identity numbers 322.4.4 Editing of a single administrative register 332.4.5 Organising the work with administrative registers 362.5 Third obstacle: The national registration system 372.5.1 Legislation governs access to data 382.5.2 Too many registers, but no good registers – what to do? 382.5.3 Legislation rules obligations to report and what to report 402.6 Why has the census been so important? 412.7 Creating the register system 422.7.1 Where do you live? 432.7.2 Where do you work? 452.7.3 With whom do you live? 462.7.4 A centralised or decentralised national system? 472.8 Register surveys and estimation methods 482.9 A traditional census or a register-based census? 49Chapter 3 The Nature of Administrative Data 513.1 Comparing questionnaire and register data 513.1.1 A questionnaire to persons compared with register data 513.1.2 An enterprise questionnaire compared with register data 543.2 Enterprise registers for combined use 563.2.1 Corrections in accounting data 573.2.2 Missing values in accounting data 583.2.3 Administrative and statistical information systems 593.3 Measurement errors in questionnaire and register data 603.3.1 Measurement errors 613.3.2 Taxation errors 62Chapter 4 Building the System – Record Linkage 654.1 Record linkage 654.2 Record linkage in the Nordic countries 664.3 Deterministic record linkage 684.4 Creating variables by adjoining and aggregation 704.5 Probabilistic record linkage 734.6 Four causes of matching errors 794.7 The statistical system and record linkage 82Chapter 5 Building the System – Quality Assessment 855.1 Four quality concepts 855.2 Making an inventory of potential sources 875.3 How can a source be used? 875.4 Quality assessment in a register-based production system 905.4.1 Analysing metadata 915.4.2 Analysis and data editing of the source 925.4.3 Comparing a source with the base register 925.4.4 Comparing a source with surveys with similar variables 935.5 Output data quality and quality of estimates 945.5.1 Analysing quality with a test census 945.5.2 Analysing quality with samples from the new register 955.5.3 Analysing quality with area samples 965.5.4 Measuring quality of basic register variables with the LFS 985.6 A coordinated system of registers 985.6.1 Are the base registers a coordinated system? 985.6.2 Quality indicators at the system level 995.7 Using the quality indicators 101Chapter 6 Building the System – Editing Register Data 1076.1 Editing in register surveys 1086.2 Editing of a single administrative register 1096.3 Consistency editing 1106.3.1 Consistency editing – is the population correct? 1116.3.2 Consistency editing – are the units correct? 1186.3.3 Consistency editing – are the variables correct? 1206.4 Case studies – editing register data 1216.4.1 Editing work within the Income and Taxation Register 1216.4.2 Editing work within the Income Statement Register 1236.4.3 What more can be learned from these examples? 1246.5 Editing, quality assessment and survey design 1256.5.1 Survey design in a register-based production system 1256.5.2 Survey design – management problems 1276.5.3 Total survey error in a register-based system 128Chapter 7 Building the System – The Population Register 1297.1 Inventory of sources 1317.1.1 Time references 1317.1.2 Activities or ‘signs of life’ 1317.2 The Population Register based on full information 1337.2.1 Object types – Changing and unchanging registers 1337.2.2 Variables with different functions in the system 1347.2.3 Updating the Population Register 1367.2.4 Registers and time 1377.2.5 Variables and time 1407.3 The Population Register in new register countries 1407.3.1 Different systems of identity numbers 1417.3.2 Problems in countries without a central Population Register 1427.3.3 How to improve coverage of the Population Register 1437.3.4 Inventory of sources – addresses and time references 1467.4 Methods to measure and improve quality 1487.4.1 Three kinds of surveys should be combined 1487.4.2 A new register-based system for statistics on persons 1507.5 Conclusions 1517.6 Challenges in old register countries 152Chapter 8 The Population Register – Estimation Methods 1558.1 Estimation in sample surveys and register surveys 1568.1.1 Estimation methods for register surveys that use weights 1578.1.2 Calibration of weights in register surveys 1578.2 Calibration of weights – the Swedish LFS 1618.2.1 Use of auxiliary information in the LFS 1618.2.2 Nonresponse bias in the LFS 1628.3 Calibration – where do people live? 1638.4 Methods to handle overcoverage 167Chapter 9 Defining Register Populations – Coverage Errors 1719.1 Defining a register’s object set 1729.1.1 Defining a population 1729.1.2 Can you alter data from the National Tax Agency? 1769.1.3 Defining a population – the Farm Register 1769.1.4 Defining a population – integrated registers 1789.2 Defining a calendar year population 1799.2.1 Defining a population – frame or register population? 1809.2.2 Sampling paradigm versus register paradigm 184Chapter 10 Building the System – The Business Register 18510.1 The Business Register and the National Accounts 18510.2 The base register for economic statistics 18710.3 The scope of the register and choice of object types 18810.3.1 The register with legal units and local units 18910.3.2 The register with enterprise units and kind of activity units 19110.4 Inventory of sources 19510.5 Creating and maintaining the Business Register 198Chapter 11 The Business Register – Estimation Methods 20111.1 Multi-valued variables 20211.2 Estimation methods 20511.2.1 Occupation in the Activity and Occupation Registers 20611.2.2 Industrial classification in the Business Register 21011.2.3 Estimates from different register versions 21311.3 Application of the method 21411.3.1 Change of industry and time series quality 21511.3.2 Transformation of weights 21711.4 A decentralised or centralised statistical system? 21811.4.1 The Calendar Year Register and the National Accounts 21911.4.2 Choosing the best source for the National Accounts 22011.5 Conclusions 224Chapter 12 Censuses, Sample Surveys and Register Surveys –Conclusions 22712.1 Attitudes towards the register-based census 22712.2 The new national statistical system 23112.2.1 The system of base registers 23212.2.2 Activity registers and longitudinal registers 23412.3 Survey design 23712.3.1 Sample survey design 23712.3.2 Register survey design 23812.3.3 Creating register variables 24112.4 Survey quality 24512.4.1 Quality of registers and register surveys 24612.4.2 The integration process – integration errors 24712.4.3 Frame errors 24712.5 Organising the new production system 24812.5.1 Enterprise architecture and the register system 24812.5.2 The register system and data warehousing 24912.5.3 Missing values – a system-based approach 25212.6 Final remarks 25412.6.1 The Statistical Population Register 25412.6.2 The system of base registers 255References 257Index 261