Advances in Business Statistics Methods and Data Collection Ger Snijkers Full Chapter

Advances in Business Statistics,
Methods and Data Collection Ger

Snijkers
Visit to download the full and correct content document:
https://ebookmass.com/product/advances-in-business-statistics-methods-and-data-co
llection-ger-snijkers/
Advances in Business Statistics, Methods and Data Collection
Advances in Business Statistics, Methods
and Data Collection
Edited by
Ger Snijkers
Statistics Netherlands
Mojca Bavdaž
University of Ljubljana
Stefan Bender
Deutsche Bundesbank and University of Mannheim
Jacqui Jones
Australian Bureau of Statistics
Steve MacFeely
World Health Organization and University College Cork
Joseph W. Sakshaug
Institute for Employment Research and Ludwig Maximilian University of Munich
Katherine J. Thompson
U.S. Census Bureau
Arnout van Delden

This edition first published 2023
© 2023 John Wiley and Sons, Inc.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted,
in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted
by law. Advice on how to obtain permission to reuse material from this title is available at
http://www.wiley.com/go/permissions.
The right of Ger Snijkers, Mojca Bavdaž, Stefan Bender, Jacqui Jones, Steve MacFeely, Joseph W. Sakshaug,
Katherine J. Thompson and Arnout van Delden to be identified as the authors of this work has been asserted in
accordance with law.
Registered Office
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
Editorial Office
111 River Street, Hoboken, NJ 07030, USA
For details of our global editorial offices, customer services, and more information about Wiley products visit us at
www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears
in standard print versions of this book may not be available in other formats.
Trademarks: Wiley and the Wiley logo are trademarks or registered trademarks of John Wiley & Sons, Inc. and/or
its affiliates in the United States and other countries and may not be used without written permission. All other
trademarks are the property of their respective owners. John Wiley & Sons, Inc. is not associated with any product
or vendor mentioned in this book.
Limit of Liability/Disclaimer of Warranty

In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow
of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review
and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment,
reagent, or device for, among other things, any changes in the instructions or indication of usage and for added
warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they
make no representations or warranties with respect to the accuracy or completeness of the contents of this work and
specifically disclaim all warranties, including without limitation any implied warranties of merchantability or
fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales
materials or promotional statements for this work. The fact that an organization, website, or product is referred to in
this work as a citation and/or potential source of further information does not mean that the publisher and authors
endorse the information or services the organization, website, or product may provide or recommendations it may
make. This work is sold with the understanding that the publisher is not engaged in rendering professional services.
The advice and strategies contained herein may not be suitable for your situation. You should consult with a
specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or
disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be
liable for any loss of profit or any other commercial damages, including but not limited to special, incidental,
consequential, or other damages.
Library of Congress Cataloging-in-Publication Data Applied for
Hardback ISBN: 9781119672302
Cover Design: Wiley

Cover Image: © Denphumi/Shutterstock
Set in 9.5/12.5pt STIXTwoText by Straive, Chennai, India

v
Contents
List of Contributors xxix
Section 1 Introduction to New Measures/Indicators for the Economy 1
1 Advances in Business Statistics, Methods and Data Collection:

Introduction 3
Ger Snijkers, Mojca Bavdaž, Stefan Bender, Jacqui Jones, Steve MacFeely,
Joseph W. Sakshaug, Katherine J. Thompson, and Arnout van Delden
1.1 The ICES-VI Edited Volume: A New Book on Establishment Statistics
Methodology 3
1.2 The Importance of Establishment Statistics 5
1.3 ICES Trends 8
1.4 Organization of This Book 11
1.4.1 Section 1: Introduction to New Measures/Indicators for the Economy 12
1.4.2 Section 2: Topics in the Production of Official Establishment Statistics and
Organizational Frameworks 12
1.4.3 Section 3: Topics in the Use of Administrative Data 13
1.4.4 Section 4: Topics in Business Survey Data Collection 14
1.4.5 Section 5: Topics in the Use of New Data Sources and New Technologies 15
1.4.6 Section 6: Topics in Sampling and Estimation 16
1.4.7 Section 7: Topics in Data Integration, Linking and Matching 16
1.5 To Conclude … 17
Disclaimer 17
References 18
Appendix: Available ICES-VI Introductory Overview Lecture (IOL) Videos 21
2 GDP and the SNA: Past and Present 23

Steve MacFeely and Peter van de Ven
2.1 Introduction 23
2.2 The Origins of National Income Statistics – A Brief History 23
2.2.1 Early Developments 23
2.2.2 Invention of Gross National Product (GNP) 24
2.2.3 The Debate on Including Government 26
2.2.4 Toward a System of National Accounts 27
vi Contents
2.2.5 Global Proliferation of GDP 27

2.3 SNA and GDP Today 28
2.3.1 The System at Large 28
2.3.2 Supply and Use Tables 28
2.3.3 Institutional Sector Accounts 31
2.3.4 The Link Between Supply and Use Tables and the Institutional Sector Accounts 33
2.3.5 Consistency and Coherence 33
2.3.6 The Relationship Between National Accounts and Business Statistics 34
2.3.6.1 Definitional Adjustments 35
2.3.6.2 Adjustments for Exhaustiveness 35
2.3.6.3 Adjustments for Time Consistency 35
2.3.6.4 Balancing Adjustments 35
2.4 Most Recent and Important Revisions to SNA (Implications for
Business Statistics) 36
2.4.1 International Standards Not Set in Stone 36
2.4.2 From SNA 1968 to SNA 1993 37
2.4.3 From SNA 1993 to SNA 2008 38
2.4.4 The SNA and Source Statistics for Enterprises 39
2.5 Conclusions and Implications for Business Statistics 40
References 41
3 GDP and the SNA: Future Challenges 43

Steve MacFeely and Peter van de Ven
3.1 Introduction 43
3.2 An Agenda for the Future 44
3.3 The Tangled Web of Globalization 44
3.4 The Digital Revolution 47
3.5 Moving Beyond GDP: GDP Impeached 48
3.6 Including a Measure of Well-being 50
3.7 Putting a Value on the Environment 52
3.8 Challenges Replacing GDP 53
3.9 Conclusions and Implications for Business Statistics 54
References 56
4 Bridging the Gap Between Business and Macroeconomic Statistics:

Methodological Considerations and Practical Solutions 63
Timo Koskimäki and Kristian Taskinen
4.1 Introduction 63
4.2 Global Production and Statistics 63
4.2.1 Concepts of Nationality and Economic Ownership 64
4.2.2 Case Finland: Global Production in Economic Statistics 65
4.2.2.1 Identification of Enterprises Involved in Global Production 65
4.2.2.2 Case on Automotive Industry 66
4.2.2.3 Foreign Trade of Goods Based on Economic Ownership 67
4.2.2.4 Challenges Related to Global Production Recordings 68
Contents vii
4.3 Co-operation Between National Statistical Offices and National Central Bank
Statistics Functions Tackling Globalization Problems 69
4.3.1 Foreign Direct Investment Network as an Example of Co-operation 69
4.3.2 Early-Warning System (EWS) 70
4.3.3 A Roadmap for Solving the Globalization-Related Issues in Monetary, Financial,
and Balance of Payments – Statistics 71
4.4 Bridging the Gap Between Business and Economic Statistics Through Global Data
Sharing 72
4.4.1 Product Innovation – One-Off or Regular Data Sharing for Better Quality 72
4.4.2 Service Innovation – Improving Respondent Service for MNEs 73
4.4.3 Process Innovation to Statistical Production by Data Sharing 73
4.4.4 Innovating User Experience – Better Relevance and Consistency for Users 74
4.4.5 Organizational Innovation – Changing the Business Model of Official Statistics 74
4.4.6 Cultural Innovation – Key to Making it Happen 75
4.4.7 Innovation in Other Industries to Learn From 75
References 76
5 Measuring Investment in Intangible Assets 79

Mojca Bavdaž, Ahmed Bounfour, Josh Martin, Alberto Nonnis, Giulio Perani,
and Tjaša Redek
5.1 Introduction 79
5.2 Data Sources on Intangibles 80
5.2.1 Past Surveys on Intangibles 81
5.2.2 Comparison of Past Surveys on Intangibles 85
5.3 Measurement Challenges in Surveys 86
5.3.1 Intangibles Are Intangible and Mobile 86
5.3.2 Own-Account Investment Prevails 87
5.3.3 Pricing of Intangibles Is Difficult 87
5.3.4 In Search of the Most Suitable Respondent for Intangibles 88
5.3.5 Investments in Intangibles Take Time 88
5.3.6 Data Existence Questioned 89
5.3.7 Evidence of Inconsistent Respondent Behavior 89
5.3.8 Summarizing the Challenges in Intangible Surveys: The 4 “F” Words 90
5.4 Intangibles and the Productivity Puzzle 91
5.4.1 Analytical Considerations 91
5.4.2 Role of Global Value Chains 91
5.5 Collecting Data on Intangibles: The Way Ahead 93
5.5.1 Methodological Improvements 93
5.5.2 Data Needs Today and Tomorrow 94
5.5.2.1 Current and Potential Users 94
5.5.2.2 A (Single) IA Survey or a Bundle of Data Sources? 95
5.5.2.3 A Parallel Development Path: Assessing Intangible Asset Stocks 95
5.6 Conclusion 96
Acknowledgment 98
References 98
viii Contents
6 Measuring the US Digital Economy 105

Jessica R. Nicholson, Thomas F. Howells III, and David B. Wasshausen
6.1 Introduction 105
6.2 Experimental Digital Economy Measures 105
6.2.1 Methodology 106
6.2.1.1 Defining the Digital Economy 107
6.2.1.2 Calculating Results 110
6.2.2 Domestic Trends 110
6.2.2.1 Value Added 110
6.2.2.2 Gross Output 112
6.2.2.3 Prices 112
6.2.3 International Collaboration and Alignment 112
6.2.3.1 The Organization for Economic Co-operation and Development Working Party on
National Accounts 113
6.2.3.2 International Comparisons 114
6.2.4 Other Areas of Research 115
6.2.4.1 “Free” Digital Media 115
6.2.4.2 Measurement and Treatment of Data 116
6.2.4.3 Prices 117
6.3 Measuring Digital Services Trade 118
6.3.1 Defining Digital Services for International Trade 119
6.3.2 Trends in ICT and ICT-enabled Services 119
6.3.3 Areas of Research 121
6.4 Conclusion and Way Forward 122
References 123
7 Establishment Based Informal Sector Statistics: An Endeavor of

Measurement from Economic Census 2018 of Nepal 125
Mahesh C. Pradhan
7.2 Issues of Informal Sector in Legislation and Policies in Nepal 125
7.2.1 Constitution 2015 126
7.2.2 Labor Act 2017 126
7.2.3 Contribution Based Social Security Act 2017 126
7.2.4 Fifteenth Periodic Plan (2019/20–2023/24) 126
7.2.5 National Employment Policy 2014 126
7.3 Concept and Definition of Informal Sector 127
7.3.1 Definition of Informal Sector from Statistical Perspective 127
7.4 Endeavors of Measuring Informal Economic Activities in Nepal 128
7.4.1 Nepal Labor Force Survey 128
7.4.2 Nepal Living Standard Surveys (NLSS) 129
7.4.3 Population Censuses 129
7.5 Economic Census 2018 130
7.5.1 Contents of Economic Census 2018 130
7.6 Status of the Informal Sector Statistics 131
7.6.1 Informal Sector Statistics from Nepal Labor Force Survey 1998 and 2008 131
7.6.2 Informal Sector Statistics from Nepal Labor Force Survey 2017/18 131
Contents ix
7.6.3 Informal Sector Statistics from National Population Census 2011 131
7.6.4 Informal Sector Statistics from National Economic Census 2018 133
7.6.5 Status of Keeping Accounting Record 133
7.6.6 Informality in Micro Small and Medium Establishments (MSME) 133
7.6.7 Street Business Situation 135
7.7 Annual Revenues/Sales, Operating Expenses in Not-Registered Establishments 137
7.8 Need of Regular Measurement Informal Sector 140
7.9 Conclusion 141
References 142
Section 2 Topics in the Production of Official Establishment Statistics

and Organizational Frameworks 145
8 Statistical Producers Challenges and Help 147

Jacqui Jones and Holly O’Byrne
8.2 A Brief Overview of the Evolution of Economic Statistics, and the Establishment of
National Statistical Institutes 147
8.3 Our Statistical Ecosystem 150
8.4 Help Available to Us 152
8.4.1 International Governance 152
8.4.2 Statistical Principles to Produce and Disseminate Official Statistics 154
8.4.3 Statistical Production Models and Frameworks 155
8.4.3.1 Quality Assurance Frameworks 156
8.4.4 Statistical Manuals and Handbooks 156
8.4.5 Classifications 156
8.4.5.1 Classifying Businesses 157
8.4.5.2 Classifying Employment and Workers 157
8.4.5.3 Classifications Overview 158
8.4.6 Statistical Tools 158
8.4.7 International Collaboration and Support 158
8.5 Summary Before the Case Study 159
8.6 Standardization Leads to Efficiency: Canada’s Integrated Business Statistics
Program 159
8.7 IBSP Objectives 160
8.8 Cornerstones of an Integrated Infrastructure System 160
8.9 Metadata-Driven Model 161
8.10 Integrated Infrastructure 161
8.11 Information Management 162
8.12 Standardization and Cooperation Within IBSP 162
8.13 The Business Register 163
8.13.1 The BR as the Common Frame 163
8.13.2 Allocation Factors on the BR 163
8.13.3 Commodities and Activities on the BR 164
8.13.4 Robust Methodologies and Generalized Systems 164
8.14 Standard Tools for Developing EQ 164
x Contents
8.15 Developing a Harmonized Content Model 165

8.16 The IBSP Data Mart and Analytical Tools 165
8.17 Managing Response Burden 166
8.18 Electronic Questionnaires 166
8.19 Large and Complex Enterprises 166
8.19.1 EPM/LAOS Programs 167
8.19.2 Customized Collection 167
8.20 Tax Replacement Strategy 167
8.21 Active Collection Management 168
8.22 Rolling Estimate Model 168
8.23 IBSP Growth and Adaptation 169
8.24 Efficiencies Gained and Learned 170
8.25 Conclusion 170
References 171
9 The Development and Maintenance of Statistical Business Registers as

Statistical Infrastructure in Statistics Indonesia and the Australian Bureau of
Statistics 175
Imam Machdi, Ratih Putri Pertiwi, Rr. Nefriana, and Willem Erasmus
9.2 The Indonesian and Australian Context 175
9.3 The Definition of a Statistical Business Register 176
9.4 The Evolution of SBRs in Statistics Indonesia and the Australian Bureau of
Statistics 176
9.4.1 Development of the Statistical Business Register in Statistics Indonesia 177
9.4.1.1 Phase 1 (1970–2012): Business Directory 177
9.4.1.2 Phase 2 (2013–2015): Integrated Business Register 177
9.4.1.3 Phase 3 (2015–2021): The Statistical Business Register 178
9.4.2 Development of the Statistical Business Register in the Australia Bureau of
Statistics 178
9.5 Statistical Business Register Designs 180
9.5.1 The Design of the BPS SBR 180
9.5.1.1 Statistical Unit Model 180
9.5.1.2 Unit Coverage 180
9.5.1.3 Data Sources 181
9.5.1.4 Main Processes 181
9.5.1.5 SBR Integration 183
9.5.2 The Design of the ABS Statistical Business Register 183
9.5.2.1 Centralized Maintenance 184
9.5.2.2 Dissemination 184
9.5.2.3 Uses 185
9.6 Statistical Business Register Benefits 185
9.7 Statistical Business Register Challenges 186
9.7.1 SBR Governance and Policy 186
9.7.2 Business Process Integration 187
Contents xi
9.7.3 System Development 187

9.8 Opportunities in SBR Implementation 188
9.8.1 Transformation Program 188
9.8.2 National Policy and Initiative 189
9.9 The Future Spine Concept 190
9.10 Conclusion 191
Acknowledgment 192
References 192
10 Managing Response Burden for Official Statistics Business

Surveys – Experiences and Recent Developments at Statistics Netherlands,
Statistics Portugal, and Statistics Sweden 193
Johan Erikson, Deirdre Giesen, Leanne Houben, and Paulo Saraiva
10.2 Understanding and Measuring Response Burden 194
10.2.1 The Concept of Response Burden 194
10.2.2 Measuring and Monitoring Response Burden 194
10.3 Organization of Response Burden Management 197
10.3.1 Legal Context and Cooperation with Other Government Bodies 197
10.3.2 Organization of Burden Management Within the NSI 198
10.4 Burden Reduction Measures 198
10.4.1 Using Alternative Sources 199
10.4.2 Improving Primary Data Collection 199
10.4.2.1 Redesigning Content to Fit Data Provision Capacities 200
10.4.2.2 Sample Coordination 200
10.4.2.3 Business Survey Communication 201
10.4.2.4 Feedback 202
10.4.2.5 File Transfer and Other Techniques (Hybrid Data Collection) 203
10.4.3 Survey-assisted Modeling with Mixed Sources 204
10.4.4 Reducing Burden Through Cooperation 204
10.4.4.1 Coordination of Metadata 205
10.4.4.2 Technical Cooperation and Standards 206
10.4.4.3 Standard Business Reporting 206
10.5 Discussion 208
Disclaimer and Acknowledgments 209
References 210
Appendix 10.A: Burden Measurement Questions 214
Appendix 10.B: Example of Statistics Netherlands’ Communication About Mandatory
Reporting 218
Appendix 10.C: Example of Adaptation Communication to Data Providers due to COVID-19
Crises – Website Statistics Netherlands 2021 219
Appendix 10.D: Personalized Feedback to Business Data Provider – Monthly Report 220
Appendix 10.E: Personalized Feedback to Business Data Provider – Yearly Report 221
Appendix 10.F: Example of RCSFI Import in Statistics Netherlands Structural Business
Statistics Questionnaire 222
xii Contents
11 Producing Official Statistics During the COVID-19 Pandemic 225

Jacqui Jones, Luisa Ryan, A.J. Lanyon, Marie Apostolou, Tanya Price, Corinna König,
Marieke Volkert, Joseph W. Sakshaug, Dane Mead, Helen Baird, Duncan Elliott,
and Craig H. McLaren
11.2 Managing the Australian Statistical Business Register During COVID-19 226
11.2.1 ABS Business Register and COVID-19 227
11.2.2 Changes to Business Reporting 228
11.2.3 Potential Impacts to the ABS BR 229
11.2.4 Increasing Number of Employers 229
11.2.5 Industry Recoding 232
11.2.6 Business Cancellations 232
11.2.7 The New Normal? 233
11.3 Mitigating COVID-19 Response Rate Risks in the Collection of the ABS Producer
and International Trade Price Indexes 234
11.3.1 Overview of ABS Producer and International Trade Price Indexes 234
11.3.2 ABS Producer and International Trade Prices Data Collection 235
11.3.3 Developing the COVID-19 Response 235
11.3.4 The Optimizing Response Strategy 236
11.3.5 Prefield Preparation 236
11.3.6 Field Development and Operations 236
11.3.7 Outcomes of the Optimizing Response Strategy 239
11.3.8 Lessons Learned 240
11.4 The Impact of Changing Data Collection Modes in the IAB Establishment Panel in
Response to the COVID-19 Pandemic 241
11.5 Classification and Statistical Implementation of Australian COVID-19 Government
Policies 247
11.5.1 Building a New Work Program 247
11.5.2 Changes to Business Surveys 248
11.5.3 Policy Case Study – JobKeeper 248
11.5.3.1 Classification of the Policy 249
11.5.3.2 Implementation of Policy Classification 249
11.5.4 Lessons for the Organization 250
11.6 Seasonal Adjustment and Trend During and Post–COVID-19 251
11.6.1 Pre-COVID Publication and Presentation of Data 252
11.6.2 Seasonal Adjustment and Trend Estimation in Practice 252
11.6.3 Publication and Presentation During COVID-19 253
11.6.4 Options for Time Series Publications During COVID-19 254
11.6.5 Modeling Outliers 254
11.6.6 Forward Factors 256
11.6.7 Option to Suspend Series 257
11.6.8 Use of High-Frequency Estimates 259
11.6.9 Other Time Series Challenges 259
11.6.10 Conclusion 259
Appendix 260
References 261
Contents xiii
Section 3 Topics in the Use of Administrative Data 265
12 Methodology for the Use of Administrative Data in Business Statistics 267

Arnout van Delden and Danni Lewis
12.2 Receive the Data 269
12.3 Inspect the Data 270
12.4 Link to Population Frame 270
12.4.1 Basic Linkage Methods 271
12.4.2 Linkage of Data Sets in the Presence of Different Unit Types 271
12.5 From Actual to Target Population 272
12.5.1 Estimation Methods to Adjust for Undercoverage 273
12.5.2 Temporary Coverage Issues 274
12.6 From Observed to Targeted Variables 274
12.6.1 Harmonization Methods 274
12.6.2 Editing Methods for Measurement Errors 276
12.6.3 Correcting for Bias Due to Decentralized and Autonomous Organizations 277
12.7 From Observed to Targeted Periods 278
12.7.1 Estimation Methods When Data Are Not Available on Time 278
12.7.1.1 Benchmarking 278
12.7.1.2 Forecasting from Previous Complete Data 279
12.7.1.3 Estimation Techniques 279
12.7.2 Estimation Methods to Adjust for Periodicity 280
12.8 Assess Data Quality 280
12.8.1 Throughput Quality 280
12.8.2 Output Quality 281
12.8.3 Analysis of Differences Between Survey and Administrative Data Estimates 282
12.9 Unsolved Issues 283
12.10 Conclusion 285
References 285
13 Developing Statistical Frameworks for Administrative Data and Integrating It

into Business Statistics. Experiences from the UK and New Zealand 291
Nicholas Cox, Craig H. McLaren, Claire Shenton, Tom Tarling, and Ella W. Davies
13.1.1 Background 291
13.1.2 Administrative Data Methods Research Program 292
13.2 Quality Frameworks for Administrative Data 293
13.2.1 Statistics Netherlands and UNECE Framework 295
13.2.2 Stats New Zealand and Zhang Framework 295
13.2.3 The ESSnet Komuso Project 296
13.2.4 The Southampton University Error Project 296
13.3 Case Study One – The Use of Value Added Tax Data in the United Kingdom 297
13.3.1 Organizational Context Within the Office for National Statistics 297
13.3.1.1 Administrative Data in Practice 298
13.3.1.2 Dealing with the Data 298
13.3.1.3 Complex Units in UK Tax Data 299
xiv Contents
13.3.2 Developing Statistical Pipelines for Processing 300

13.3.3 The Use of Administrative Data in UK Monthly Short-term Indicators 301
13.3.4 The Use of Administrative Data for Regional Estimation 301
13.3.5 Example: Comparison of VAT Data to Survey Data 303
13.4 Case Study 2 – A Greater Use of Administrative Data in New Zealand’s Labor Market
Statistics 305
13.4.1 Organizational Context 305
13.4.2 Stats NZ’s New Monthly Employment Indicator 307
13.4.3 Redesigning the Quarterly Employment Survey 308
13.4.3.1 Phase One 309
13.4.3.2 Phase Two 310
13.4.4 Introducing Stats NZ’s New Quarterly Business Employment Data 310
13.5 Concluding Remarks 311
References 311
14 The Evolution of Integrating Administrative Data in Business Statistics in

Ireland 315
Colin Hanley and Sorcha O’Callaghan
14.2 Administrative Data 316
14.2.1 Benefits 316
14.2.1.1 Resources 317
14.2.1.2 Coverage 317
14.2.1.3 Timeliness 317
14.2.1.4 Response Burden 317
14.2.2 Challenges 318
14.2.2.1 Access 318
14.2.2.2 Quality 318
14.2.2.3 Statistical Units 318
14.3 Administrative Data in CSO Business Statistics 319
14.3.1 Legal Mandate 319
14.3.2 Business Register 319
14.3.3 Structural Business Statistics (SBS) 321
14.4 Data Linkage Using Administrative Data 322
14.4.1 Exporting Enterprises in Ireland 322
14.4.1.1 Trade Data Sources 322
14.4.1.2 Linking 323
14.4.1.3 Linking Trade to Business Register 323
14.4.1.4 Results 324
14.4.2 Business Signs of Life 325
14.4.2.1 Linking 326
14.4.2.2 Results 327
14.4.3 Lessons Learned 329
14.4.3.1 Quality 329
14.4.3.2 Profiling 329
14.4.3.3 Coverage 329
14.4.3.4 Transparency 329
Contents xv
14.5 The Use of VAT Data in Business Statistics 330

14.5.1 The Current Situation 330
14.5.2 VAT Data Available in Ireland 330
14.5.3 VAT Data for Short-Term Business Statistics 330
14.5.4 VAT Data as a Timely Indicator of Business Signs of Life 331
14.5.5 Business Signs of Life Series 331
14.6 Summary 331
References 332
Section 4 Topics in Business Survey Data Collection 335
15 What Computerized Business Questionnaires and Questionnaire

Management Tools Can Offer 337
Gustav Haraldsen
15.2 Business Survey Challenges 338
15.2.1 Concepts 338
15.2.2 Units 339
15.2.3 Time References 340
15.3 The Path to Competent and Motivated Respondents 341
15.4 What Computerization Can Offer 343
15.4.1 Source-Oriented Instruments 343
15.4.2 Combined Communication Means 345
15.4.3 Designed Dialogues 348
15.5 What We Know That We Don’t Know 351
Acknowledgments 352
References 352
16 Tailoring the Design of a New Combined Business Survey: Process, Methods,

and Lessons Learned 357
Ger Snijkers, Leanne Houben, and Fred Demollin
16.2 Toward the New “CBS-DNB Finances of Enterprises and Balance of Payments”
Survey 359
16.3 Achieving Coherent Statistics 359
16.4 Questionnaire Communication: Tailoring the Design 361
16.4.1 Steps in the Questionnaire Development Process 362
16.4.2 The Feasibility Study 363
16.4.3 What Data and Where Are the Data? 365
16.4.4 The Business Response Process 366
16.4.5 How: Questionnaire Design Requirements 367
16.4.6 Questionnaire Development 368
16.5 Introducing the Survey in the Field: The Survey Communication Strategy 370
16.5.1 The Three Phases in the Survey Communication Strategy 370
16.5.1.1 Pre-field Phase 370
16.5.1.2 Field Phase of the First Wave of the Survey 372
xvi Contents
16.5.1.3 Post-field Phase of the First Wave of the Survey 373

16.5.2 Evaluation of the Communication Strategy: Was the Strategy Effective? 374
16.5.2.1 Effectiveness of the Pre-field Strategy 374
16.5.2.2 The Response Rate Development in the Next Quarters of 2019 378
16.6 Lessons Learned 379
Acknowledgment 381
Disclaimer 381
References 382
17 Advances in Question(naire) Development, Pretesting, and Evaluation 387

Diane K. Willimack, Heather Ridolfo, Amy Anderson Riemer, Melissa Cidade, and Kathy Ott
17.2 Adaptation and Innovation in Pretesting Methods 388
17.2.1 Case Study #1: Emerging Topic of Robotics – it takes a village 389
17.2.1.1 Background 389
17.2.1.2 Pretesting Methodology 389
17.2.1.3 Summary 391
17.2.2 Case Study #2: Multiple Modes and Methods – different strokes for different folks 391
17.2.2.3 Summary 393
17.2.3 Case Study #3: Record-Keeping Study – “I don’t keep my records that way” 394
17.2.3.2 Pretesting Methodology for Phase I 394
17.2.3.3 Pretesting Methodology for Phase II 395
17.2.4 Case Study #4: Usability Testing – when human meets computer 397
17.2.4.3 Summary 401
17.2.5 Case Study #5: Remote Testing, Logistics, and COVID – reality is virtual 401
17.2.5.2 Remote Testing at NASS 401
17.2.5.3 Remote Testing at the Census Bureau 401
17.2.5.4 Summary 403
17.2.6 Case Study #6: Pretesting Plus Paradata – a look underneath the hood 403
17.3 Pretesting Methodologies: Current Features and Future Needs 405
17.3.1 Current Features and Consequences 405
17.3.1.1 Finding “the Missing Link”: Collaborative Partnerships 405
17.3.1.2 The Odyssey: Exploratory Methods 405
17.3.1.3 Leave No Stone Unturned: Multiple Methods 405
17.3.1.4 Methuselah: The Many Roles of Technology 406
17.3.1.5 Trust the Process: The Response Process 406
17.3.2 Future Needs and Implications 406
Acknowledgments 408
References 408
Contents xvii
18 Using Paradata in Electronic Business Survey Questionnaires 413

Ger Snijkers, Susan Demedash, and Jessica Andrews
18.2 Paradata 414
18.3 Questionnaire Completion Paradata 417
18.4 Looking Inside the Questionnaire Completion Process 419
18.4.1 Completing the CBS-DNB Quarterly Survey on Finances of Enterprises and Balance
of Payments 421
18.4.1.1 Questionnaire Completion Profiles 421
18.4.1.2 Usage of Download and Import Functions 423
18.4.1.3 Usage of Dutch and English Versions of the Questionnaire 424
18.4.1.4 Time Needed to Complete the Questionnaire 424
18.4.1.5 Effect of the Communication Strategy 426
18.4.2 Business Survey Use of Paradata at Statistics Canada 427
18.5 Conclusions 431
Acknowledgment 432
Disclaimer 433
References 433
19 Recent Findings from Experiments in Establishment Surveys 437

Josh Langeland, Heather Ridolfo, Jaki McCarthy, Kathy Ott, Doug Kilburg, Karen CyBulski,
Melissa Krakowiecki, Larry Vittoriano, Matt Potts, Benjamin Küfner, Joseph W. Sakshaug,
and Stefan Zins
19.2 Experiments with Mailed Survey Packets to Improve Recruitment Strategies in a
National Establishment Survey (BLS) 438
19.2.1 Motivation 438
19.2.2 Advance Letter Study 439
19.2.2.1 Experiment Design 439
19.2.2.2 Results 440
19.2.3 Folder Design Study 441
19.2.3.1 Experiment Design 441
19.2.3.2 Results 441
19.2.4 Discussion 443
19.3 Experiments Testing Changes to Data Collection Timing and Content of Contacts in the
US Census of Agriculture (NASS) 443
19.4 Comparing FedEx to Traditional Postage in a Survey of Substance Abuse and Mental
Health Facilities (by Mathematica for SAMHSA) 451
19.4.2 Experiment Design 452
19.4.3 Results 454
19.5 Addressing Item Nonresponse with Clarifying Information – Evidence from the IAB
Job Vacancy Survey (IAB) 458
19.5.2 Experiment Design 459
xviii Contents
19.5.3 Results 461

19.5.3.1 Item Duration 461
19.5.3.2 Item Nonresponse 461
19.5.3.3 Spillover Effects 462
19.6 Summary 464
Acknowledgment 465
References 465
20 Web Portals for Business Data Collection 469

Bente Hole and Leanne Houben
20.2 The NSI Web-Portal Study 470
20.2.1 More About the Survey 470
20.2.2 Survey Results and Other Findings 471
20.2.2.1 Type and Size of Portals 472
20.2.2.2 Common Features and Status 474
20.2.2.3 Registration, Authentication, and Authorization 476
20.2.2.4 Data Import and Transfer 481
20.2.2.5 Identifying the Right Business Unit 483
20.2.2.6 Returning Data to the Respondents 484
20.2.2.7 Contact Options and Communication 485
20.2.2.8 Strengths and Weaknesses 487
20.3 Investigating How to Build a Customized Portal at Statistics Netherlands 488
20.4 Recommendations and Future Developments 489
20.4.1 Recommendations 489
20.4.2 Future Web Portal Developments 491
Disclaimer and Acknowledgements 493
Appendix 20.A List of NSIs the Questionnaire was Sent to 493
Appendix 20.B Word Copy of Questionnaire 494
References 498
21 A Creative Approach to Promoting Survey Response 501

Charles F. Brady, Jr. and Kari L. Klinedinst
21.2 Background 502
21.3 Approach and Methods 503
21.3.1 Public Sector – Private Sector Partnership 503
21.3.2 Strategic Objectives 504
21.3.3 Target Segmentation and Focus Groups 505
21.3.4 Communications Plan 506
21.4 Results from Focus Groups 507
21.4.1 Focus Groups – Round One 507
21.4.2 Focus Groups – Round Two 508
21.4.3 Intermediary Meeting at Census 508
21.4.4 Focus Groups Round Three 509
Contents xix
21.4.5 Additional Meetings 509

21.5 Development of Campaign Materials 510
21.5.1 Brochures 511
21.5.2 Videos 512
21.5.3 Campaign Website 512
21.5.4 Partner Briefing Presentation Content 513
21.5.5 Island Areas 513
21.6 Campaign Implementation 513
21.6.1 “9-8-7” Campaign 513
21.6.2 Webinar 514
21.6.3 Email Awareness Campaign 514
21.6.4 Internal Communications 514
21.6.5 Economic Census Day for Census Bureau Staff 514
21.6.6 Media Relations 514
21.6.7 Congressional and Intergovernmental Affairs 514
21.6.8 Social Media 515
21.6.9 Meetings and Events 515
21.7 Moving Forward 515
21.7.1 Respondent Portal Changes 515
21.7.2 Adaptation for Current Surveys 516
21.8 Conclusion 516
References 517
Section 5 Topics in the Use of New Data Sources and New

Technologies 519
22 Statistical Data Production in a Digitized Age: The Need to Establish

Successful Workflows for Micro Data Access 521
Stefan Bender, Jannick Blaschke, and Christian Hirsch
22.2 Building Blocks for Successful Workflows Enabling Access to Micro Data 522
22.2.1 Building Block 1: Laying the Technical and Procedural Foundations 523
22.2.2 Building Block 2: Generating Safe Results 525
22.2.3 Building Block 3: Generating Value for All Stakeholders 526
22.3 An Alternative Approach to Measuring Value: FAIR Data 527
22.4 Applying the BUBMIC Model to Research Data Centers 528
22.4.1 Building Block 1: Laying the Technical and Procedural Foundations 529
22.4.2 Building Block 2: Generating Safe Results 530
22.4.3 Building Block 3: Generating Value for All Stakeholders 531
22.4.4 Examples of Generating Value for All Stakeholders 531
22.4.4.1 Rules for Visiting Researchers at the RDSC 531
22.4.4.2 SDC Packages in Stata and R 532
22.4.4.3 Dobby, the BBk-RDSC’s High-Performance and Streamlined Data Production
Pipeline 532
xx Contents
22.4.4.4 RDSC Contract Generator 532

22.4.4.5 Annodata Schema 532
22.5 Conclusion 534
Acknowledgments 534
Disclaimer 534
References 534
23 Machine Learning in German Official Statistics 537

Florian Dumpert
23.2 Terminology and a Short Introduction to Machine Learning 537
23.3 Machine Learning in Official Statistics – International Overview 541
23.4 History and Current Status of Machine Learning in German Official Statistics 543
23.4.1 Federal Statistical Office of Germany 543
23.4.2 History and Current Status in the German Official Statistics Network 544
23.5 Some Current Projects at the Federal Statistical Office 545
23.5.1 Overview 545
23.5.2 Machine Learning to Increase Analysis Capabilities in the Area of Minimum Wage
Using Official Statistics 545
23.5.3 Machine Learning for Editing and Imputation 550
23.5.3.1 Relevance 550
23.5.3.2 Editing and Imputation in the New Digital Earnings Survey 550
23.5.3.3 Studies on the Preservation of the Distribution under Imputation 552
23.6 Summary and Outlook 555
References 557
24 Six Years of Machine Learning in the Bureau of Labor Statistics 561

Alexander Measure
24.2 Why Official Statistics? 561
24.3 How Should We Do It? 562
24.4 Is It Good Enough? 564
24.5 How Should We Use It? 565
24.6 How Do We Integrate It? 567
24.7 How Do We Maintain It? 568
24.8 Conclusion 570
References 570
Appendix: F1-score and Macro F1-score 572
25 Using Machine Learning to Classify Products for the Commodity Flow

Survey 573
Christian Moscardi and Benjamin Schultz
25.1 Background 573
25.1.1 Commodity Flow Survey (CFS) Background 573
25.1.2 CFS Data Collection Challenges 573
Contents xxi
25.1.2.1 Nonrespose 573

25.1.2.2 Data Quality 574
25.1.2.3 Respondent Burden 574
25.1.2.4 Related Work 575
25.2 Data 575
25.3 Methods 577
25.3.1 Filtering and Text Cleansing 577
25.3.1.1 Filtering CFS Response Data 577
25.3.1.2 Text Preprocessing 578
25.3.1.3 De-duplication and Disambiguation 578
25.3.2 Deriving Variables (Features) from Text Data 580
25.3.3 Other Features Incorporated into Model – NAICS Code 580
25.3.4 Resolving Previously Unseen Variables during Prediction 581
25.3.5 Model 581
25.3.6 Training and Evaluation 582
25.3.7 Imputing and Correcting Data – Edit Rule Agreement 583
25.4 Results 584
25.4.1 Model Results 584
25.4.2 Applications to 2017 CFS and Impact of ML 586
25.4.2.1 Applications 586
25.4.2.2 Impact of ML 587
25.5 Conclusion and Future Work 588
Disclaimer 590
References 590
26 Alternative Data Sources in the Census Bureau’s Monthly State Retail Sales
Data Product 593
Rebecca Hutchinson, Scott Scheleur, and Deanna Weidenhamer
26.1 Introduction/Overview 593
26.2 History of State-Level Retail Sales at Census 594
26.3 Overview of the MSRS 595
26.4 Methodology 597
26.4.1 Directly Collected Data Inputs 597
26.4.2 Frame Creation 598
26.4.3 Estimation and Imputation 599
26.4.3.1 Composite Estimator 599
26.4.3.2 Synthetic Estimator 600
26.4.3.3 Hybrid Estimator 601
26.4.4 Quality Metrics 602
26.5 Use of Alternative Data Sources in MSRS 602
26.5.1 Input to MSRS Model 602
26.5.2 Validation 605
26.6 Conclusion 609
Disclaimer 610
References 610
xxii Contents
Section 6 Topics in Sampling and Estimation 613
27 Introduction to Sampling and Estimation for Business Surveys 615

Paul A. Smith and Wesley Yung
27.2 Statistical Business Registers 615
27.3 Sampling 619
27.3.1 Stratified Sampling 619
27.3.2 Cut-off Sampling 622
27.3.3 Probability Proportional to Size Sampling 622
27.3.4 Indirect Sampling 623
27.3.5 Balanced Sampling 624
27.4 Estimation 624
27.4.1 Model-assisted and Calibration Estimation 624
27.4.2 Outliers 628
27.5 Model-based Estimation 629
27.5.1 Small Area Estimation 629
27.5.2 Nowcasting 630
27.5.3 Model-based Estimators 631
27.6 Conclusion 631
References 632
28 Sample Coordination Methods and Systems for Establishment Surveys 637

Alina Matei and Paul A. Smith
28.2 Sample Coordination 638
28.2.1 Notation and Definitions 638
28.2.2 Methods for Sample Coordination 638
28.2.3 Methods Based on PRNs 639
28.2.4 Non-PRN Methods 642
28.3 Comparing Sample Coordination Methods 644
28.3.1 Measures Used in Sample Coordination 644
28.3.2 Criteria for Sample Coordination 645
28.4 Sample Coordination Systems 647
28.4.1 Optimization Measures in Sample Coordination Systems 648
28.5 Overview of Sample Coordination Systems 648
28.5.1 Coordinated Poisson Sampling/Conditional Selection 648
28.5.2 SAMU 650
28.5.3 Synchronized Sampling 650
28.5.4 Burden-Based Coordination 651
28.5.5 Coordination Functions 651
28.6 Discussion 652
28.6.1 Distinguishing Sample Coordination Methods and Sample Coordination Systems 652
28.6.2 Further Challenges 653
28.7 Conclusion 653
Acknowledgments 654
References 654
Contents xxiii
29 Variance Estimation for Probability and Nonprobability Establishment

Surveys: An Overview 659
Jill A. Dever and Dan Liao
29.1 Estimation for Probability Business Survey Data 660
29.1.1 Probability Sampling in Practice 660
29.1.2 Theories of Population Inference 660
29.1.3 Basic Weighting Steps 662
29.1.4 Variance Estimation for Probability Surveys 662
29.1.4.1 Exact Formulas 663
29.1.4.2 Linearization Methods 664
29.1.4.3 Replication Methods 665
29.1.5 Variance Estimation with Imputed Values 671
29.1.6 Variance Estimation Applications Among Probability Establishment Surveys 671
29.2 Estimation with Nonprobability Establishment Survey Data 673
29.2.1 Nonprobability Sampling in Practice 673
29.2.2 Analytic Objectives 673
29.2.2.1 Methods for Nonprobability Estimation 675
29.2.2.2 Methods for Hybrid Estimation 676
29.2.3 Variance Estimators 677
29.2.3.1 Quasi-randomization Methods 677
29.2.3.2 Superpopulation and Model-based Methods 678
29.3 Concluding Remarks 678
References 679
30 Bayesian Methods Applied to Small Area Estimation for Establishment

Statistics 685
Paul A. Parker, Ryan Janicki, and Scott H. Holan
30.2 Bayesian Hierarchical Modeling for Dependent Data 688
30.3 Area-Level Models 690
30.4 Unit-Level Models 692
30.4.1 Basic Unit-Level Model 692
30.4.2 Accounting for Survey Design 693
30.4.3 Models for Non-Gaussian Data 694
30.5 Empirical Simulation Study 695
30.6 Data Analysis 698
30.7 Discussion 701
Acknowledgments 702
References 702
31 Variance Estimation Under Nearest Neighbor Ratio Hot Deck Imputation for
Multinomial Data: Two Approaches Applied to the Service Annual Survey
(SAS) 705
Rebecca Andridge, Jae Kwang Kim, and Katherine J. Thompson
31.2 Basic Setup 709
31.3 Single Imputation Variance Estimation 710
xxiv Contents
31.4 Multiple Imputation Variance Estimation 711

31.5 Simulation Study 712
31.5.1 Data Generation 712
31.5.2 Imputation Methods Implemented 714
31.5.3 Evaluation of Performance 715
31.5.4 Results 716
31.6 Empirical Application 718
31.6.1 Background 718
31.6.2 Results 720
31.7 General Conclusion 722
Acknowledgments 724
Disclaimer 724
References 724
32 Minimizing Revisions for a Monthly Economic Indicator 727

Nicole Czaplicki, Stephen Kaputa, and Laura Bechtel
32.2 MARTS and MRTS Background and Motivation 729
32.2.1 Sample Design 729
32.2.2 Unit Definitions 730
32.2.3 Response Rates 730
32.2.4 Estimation Methodology 730
32.2.5 Imputation Methodology and Procedures 732
32.3 Estimation Evaluation 733
32.3.1 Estimation Methods Considered 733
32.3.1.1 Link Relative Estimation 733
32.3.1.2 Modified Link Relative Estimator 734
32.3.1.3 Weighting Class Estimator 734
32.3.1.4 Ratio Estimator 735
32.3.2 Estimation Evaluation Criteria 735
32.3.3 Estimation Empirical Results 736
32.3.4 Estimation Discussion 737
32.4 Automating the Detection of High-Priority Units for Imputation 738
32.4.1 Methods for Identifying Units for Analyst Imputation 738
32.4.1.1 Influence Measure Method (Month-to-Month Change) 738
32.4.1.2 Size Identification Method 739
32.4.1.3 Prioritization 741
32.4.2 Identifying High-Priority Units for Analyst Imputation Evaluation 742
32.4.3 Identifying High-Priority Units for Analyst Imputation Discussion 742
32.5 Automating Imputation Procedures 743
32.5.1 Imputation Model 743
32.5.1.1 RegARIMA Time Series Model 744
32.5.1.2 Hierarchical Bayesian Regression (HBR) Model 745
32.5.2 Simulation Study 745
32.5.2.1 Simulation Study Design 746
32.5.2.2 Evaluation Statistics 746
32.5.2.3 Simulation Results 747
Contents xxv
32.5.3 Imputation Discussion 750

32.6 Conclusion 751
Acknowledgments 752
Disclaimer 752
References 752
Section 7 Topics in Data Integration, Linking and Matching 755
33 Record Linkage for Establishments: Background, Challenges, and an

Example 757
Michael D. Larsen and Alan Herning
33.2 Variables for Linking Records 758
33.3 Exact, Deterministic, and Probabilistic Matching 759
33.3.1 Exact Matching and Deterministic Matching with Multiple Passes 759
33.3.2 Probabilistic Matching 761
33.3.3 Combining Deterministic and Probabilistic Linkage 763
33.4 Additional Considerations in Record Linkage 763
33.4.1 Structural Considerations in Record Linkage 763
33.4.2 One-to-One Matching 764
33.4.3 Using Linked Files in Analysis 765
33.4.4 Confidentiality and Computing 766
33.5 A Practical Example of Business Data Linking: The Business Longitudinal Analysis
Data Environment 768
33.5.1 Overview 768
33.6 BLADE Linking Methodology 768
33.6.1 BLADE Integrating Frame 768
33.6.2 Linking Data to BLADE 770
33.6.2.1 ABS Survey Data 770
33.6.2.2 BLADE Core 770
33.6.2.3 Linkage of Other Administrative Data 771
33.6.3 Maintaining BLADE 772
33.6.3.1 Static and Dynamic BLADE 773
33.7 BLADE Access Model 773
33.7.1 Accessing BLADE 774
33.7.2 Customized BLADE Products 775
33.7.3 How BLADE Is Being Used 775
33.8 Conclusion 775
References 776
34 Methods for Estimating the Quality of Multisource Statistics 781

Arnout van Delden, Sander Scholtus, Ton de Waal, and Irene Csorba
34.2 Representation Error 783
34.2.1 Estimating and Correcting Selectivity 783
34.2.1.1 Estimating Output Accuracy with Respect to Representation Errors 783
xxvi Contents
34.2.1.2 Correcting for Representation Errors 785

34.2.2 Case Study on Estimating Bias Due to Selectivity 785
34.2.3 Case Study on Correcting Output for Selectivity 787
34.3 Linkage Error 788
34.3.1 Estimating the Effect of Linkage Errors on Outputs 788
34.3.2 Adjusting Outputs to Correct for Linkage Errors 790
34.3.3 Simulation Study on Correcting Contingency Tables for Linkage Errors 792
34.4 Measurement Error 793
34.4.1 Quantifying Measurement Error 794
34.4.2 Estimating the Effect of Measurement Error on Outputs 795
34.4.3 Case Study on Turnover Growth Rates by Economic Activity 796
34.4.4 Case Study on Internet Purchases 797
34.4.5 Obtaining a Bias-corrected Estimator for Measurement Errors 798
34.4.6 Simulation Study Comparing the Bias-corrected Estimators 799
34.5 Conclusion 800
References 801
35 Adopting Previously Reported Data into the 2022 Census of Agriculture:

Lessons Learned from the 2020 September Agricultural Survey 805
Linda J. Young, Joseph B. Rodhouse, Zachary Terner, and Gavin Corral
35.2 Agricultural Survey and Previously Reported Data (PRD) 806
35.3 Study Design 807
35.3.1 Survey Questions and PRD Included in the Study 807
35.3.2 Design of the Study 808
35.3.3 Data Collection 810
35.3.4 Analysis of the Study 810
35.4 Study Results 810
35.4.1 Comparison of Response Rates 811
35.4.2 Comparison of Completion Times 812
35.4.3 Analysis of Update Rates 814
35.4.4 Analysis of Edit Rates 816
35.5 Discussion 816
Acknowledgments 818
References 819
36 Integrating Alternative and Administrative Data into the Monthly Business

Statistics: Some Applications from Statistics Canada 821
Marie-Claude Duval, Richard Laroche, and Sébastien Landry
36.1 Context for Integrating Alternative and Administrative Data 821
36.2 Replacement of Survey Data by Tax Data in the Monthly Survey of Food Services and
Drinking Places 822
36.2.1 Previous Methodology Under a Survey Design 822
36.2.2 The GST Administrative File 823
36.2.3 The New Methodology Proposed to Replace Survey Data with Administrative
Data 824
36.2.4 Assessment and Requirements Before the Implementation 825
Contents xxvii
36.2.5 Implementation 825

36.2.6 Conclusion and Future Work 826
36.3 Replacement of Survey Data by Scanner Data in the Retail Commodity Survey 827
36.3.1 The Retail Commodity Survey 827
36.3.2 Scanner Data File and Coding by Commodity 828
36.3.3 Methodology to Replace Survey Data with Scanner Data 829
36.3.4 Quality Assurance Process 829
36.3.5 Future Work and Conclusion 830
36.4 Integration of Survey Data with Administrative Data in the Monthly Survey of
Manufacturing Using Small Area Estimation Techniques 831
36.4.1 Methodology of the Monthly Survey of Manufacturing (MSM) 831
36.4.2 Small Area Estimation (SAE) Method and Implementation 831
36.4.3 Confidentiality Strategy for the SAE Estimates 833
36.4.4 Conclusion and Future Work 834
36.5 Impact of the COVID-19 Pandemic in the Monthly Business Programs and the
Administrative and Alternative Data Sources Used 834
36.5.1 Survey Responses 834
36.5.2 Imputation Methods 835
36.5.3 Scanner Data and Administrative GST Data 835
36.5.4 Advance Estimates and Supplementary Questions 836
36.6 Future Initiatives for the Monthly Business Programs 836
36.7 Conclusion 837
Acknowledgments 837
References 838
Index 839
k
Downloaded from https://onlinelibrary.wiley.com/doi/ by Lao People's Dem Rep Hinari access, Wiley Online Library on [09/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
xxix
List of Contributors
Jessica Andrews Laura Bechtel

Statistics Canada Economic Statistical Methods Division
Ottawa US Census Bureau
Ontario Washington
Canada DC
USA
Rebecca Andridge
Division of Biostatistics Stefan Bender
The Ohio State University Data Service Center
Columbus Deutsche Bundesbank
k OH Frankfurt k
USA Germany
Marie Apostolou and

Melbourne School of Social Science
Victoria University of Mannheim
Australia Mannheim
Germany
Helen Baird
Australian Bureau of Statistics Jannick Blaschke
Hobart Data Service Center
Tasmania Deutsche Bundesbank
Australia Frankfurt
Germany
Mojca Bavdaž
Academic Unit for Mathematics, Statistics, and Ahmed Bounfour
Operations Research RITM & European Chair on Intangibles
School of Economics and Business Université Paris-Saclay
University of Ljubljana Sceaux
Ljubljana France
Slovenia
k
k
xxx List of Contributors
Charles F. Brady, Jr. Susan Demedash

US Census Bureau Statistics Canada
Washington Ottawa
DC Ontario
USA Canada
Melissa Cidade Fred Demollin

U.S. Census Bureau Department of Business Statistics (Heerlen)
Washington Division of Economic and Business Statistics
DC and National Accounts
USA Statistics Netherlands
Heerlen
Gavin Corral The Netherlands
USDA National Agricultural Statistics Service
Washington Jill A. Dever
DC RTI International
USA Washington
DC
Nicholas Cox USA
Statistics NZ
Wellington Florian Dumpert
New Zealand Federal Statistical Office of Germany
Wiesbaden
k Irene Csorba
k
Germany
Vrije Universiteit Amsterdam
Amsterdam Marie-Claude Duval
The Netherlands Statistics Canada
Ottawa
Karen CyBulski ON
Mathematica Canada
Princeton
NJ Duncan Elliott
USA Office for National Statistics
Newport
Nicole Czaplicki Wales
Economic Statistical Methods Division UK
US Census Bureau
Washington Willem Erasmus
DC ABS – Australian Bureau of Statistics
USA Business Register Unit
Melbourne
Arnout van Delden Australia
Department of Research and Development
(Methodology)
Heerlen/the Hague
The Netherlands
k
k
List of Contributors xxxi
Johan Erikson and

Statistics Sweden
Örebro Office of the Associate Director for Research
Sweden and Methodology
U.S. Census Bureau
Deirdre Giesen Washington
Statistics Netherlands DC
Heerlen USA
The Netherlands
Bente Hole
Colin Hanley Statistics Norway
Central Statistics Office Kongsvinger
Cork City Norway
Ireland
Leanne Houben
Gustav Haraldsen Department of Innovation
Division for Methods Development and Functional Management
Statistics Norway (Innovation Data Services)
Oslo/Kongsvinger Division of Data Services, Research and
Norway Innovation, Statistics Netherlands
Heerlen
Alan Herning The Netherlands
Data Strategy, Integration and Services
k Division Thomas F. Howells III k
Australian Bureau of Statistics Chief, Industry Economics Division
Canberra Bureau of Economic Analysis
Australia USA
Christian Hirsch Rebecca Hutchinson

Data Service Center US Census Bureau
Deutsche Bundesbank Washington
Frankfurt DC
Germany USA
Scott H. Holan Jacqui Jones

Department of Statistics Macroeconomic Statistics Division
University of Missouri Australian Bureau of Statistics
Columbia Canberra
MO ACT
USA Australia
Ryan Janicki
Center for Statistical Research and
Methodology
U. S. Census Bureau
Washington
DC
USA
k
k
xxxii List of Contributors
Stephen Kaputa Sébastien Landry

Economic Statistical Methods Division Statistics Canada
US Census Bureau Ottawa
Washington ON
DC Canada
USA
Josh Langeland
Doug Kilburg U.S. Bureau of Labor Statistics
U.S. National Agricultural Statistics Service Washington
Washington DC
DC USA
USA
A.J. Lanyon
Jae Kwang Kim Australian Bureau of Statistics
Department of Statistics Brisbane
Iowa State University Queensland
Ames Australia
IA
USA Richard Laroche
Statistics Canada
Kari L. Klinedinst Ottawa
US Census Bureau ON
Washington Canada
k k
DC
USA Michael D. Larsen
Department of Mathematics and Statistics
Corinna König Saint Michael’s College
Institute for Employment Research Colchester
Nuremberg VT
Germany USA
Timo Koskimäki Danni Lewis

Statistics Finland Office for National Statistics
Helsinki Newport
Finland UK
Melissa Krakowiecki Dan Liao

Mathematica RTI International
Princeton Washington
NJ DC
USA USA
Benjamin Küfner Steve MacFeely

German Institute for Employment Research Director of Data and Analytics
Nuremberg Department of Data and Analytics
Germany World Health Organization
Geneva
Switzerland
k
k
List of Contributors xxxiii
Imam Machdi Christian Moscardi

BPS – Statistics Indonesia U.S. Census Bureau
Deputy of Methodology and Statistical Economic Reimbursable Surveys Division
Information Business Development Staff
Jakarta Washington
Indonesia DC
USA
Josh Martin
Office for National Statistics Rr. Nefriana
Newport BPS – Statistics Indonesia
UK Deputy of Methodology and Statistical
Information
Alina Matei Jakarta
Institute of Statistics Indonesia
University of Neuchâtel
Neuchâtel Jessica R. Nicholson
Switzerland Research Economist
Bureau of Economic Analysis
Jaki McCarthy USA
U.S. National Agricultural Statistics Service
Washington Alberto Nonnis
DC RITM & European Chair on Intangibles
USA Université Paris-Saclay
k k
Sceaux
Craig H. McLaren France
Office for National Statistics
Newport Holly O’Byrne
Wales Statistics Canada
UK Ottawa
Ontario
Dane Mead Canada
Brisbane Sorcha O’Callaghan
Queensland Central Statistics Office
Australia Cork City
Ireland
Alexander Measure
U.S. Bureau of Labor Statistics Kathy Ott
Washington USDA/National Agricultural Statistics Service
DC Washington
USA DC
USA
k
k
xxxiv List of Contributors
Paul A. Parker Heather Ridolfo

Department of Statistics U.S. Energy Information Administration
University of Missouri Washington
Columbia DC
MO USA
USA
Amy Anderson Riemer
Ratih Putri Pertiwi U.S. Census Bureau
BPS – Statistics Indonesia Washington
Deputy of Methodology and Statistical DC
Information USA
Jakarta
Indonesia Joseph B. Rodhouse
USDA National Agricultural Statistics Service
Giulio Perani
Washington
ISTAT
DC
Rome
USA
Italy
Matt Potts Luisa Ryan

Mathematica Australian Bureau of Statistics
Princeton Melbourne
NJ Victoria
USA Australia
k k
Mahesh C. Pradhan Tom Tarling

Economic Census Section Office for National Statistics
Central Bureau of Statistics Newport
Kathmandu Wales
Nepal UK
Tanya Price Joseph W. Sakshaug

Australian Bureau of Statistics German Institute for Employment Research
Canberra Nuremberg
ACT Germany
Australia and
Ludwig Maximilian University of Munich
Tjaša Redek
Department of Statistics
School of Economics and Business
Munich
University of Ljubljana
Germany
Ljubljana
Slovenia
and
Heather Ridolfo
University of Mannheim
U.S. Energy Information Administration
Department of Sociology
Washington
Mannheim
DC
Germany
USA
k
k
List of Contributors xxxv
Joseph W. Sakshau Ger Snijkers

Statistical Methods Research Department Department of Research and Development
Institute for Employment Research (Methodology)
Nuremberg Division of Data Services, Research and
Germany Innovation, Statistics Netherlands
Heerlen
Paulo Saraiva The Netherlands
Statistics Portugal
Lisbon Kristian Taskinen
Portugal Statistics Finland
Helsinki
Scott Scheleur Finland
US Census Bureau
Washington Zachary Terner
DC USDA National Agricultural Statistics Service
USA Washington
DC
Sander Scholtus USA
The Hague and
The Netherlands
National Institute of Statistical Sciences
Benjamin Schultz Washington
k U.S. Census Bureau DC
k
Economic Management Division USA
Washington
DC Katherine J. Thompson
USA Associate Directorate for Economic Programs
U.S. Census Bureau
Claire Shenton
Washington
Office for National Statistics
DC
Newport
USA
Wales
UK Peter van de Ven
Inter-Secretariat Working Group on National
and
Accounts
Department of Statistics New York
Ludwig Maximilian University of Munich USA
Munich
Germany Larry Vittoriano
Mathematica
Paul A. Smith Princeton
S3RI and Department of Social Statistics & NJ
Demography USA
University of Southampton
Southampton
UK
k
k
xxxvi List of Contributors
Marieke Volkert Diane K. Willimack

Institute for Employment Research U.S. Census Bureau
Nuremberg Washington
Germany DC
USA
Ton de Waal
Statistics Netherlands Linda J. Young
The Hague USDA National Agricultural Statistics Service
The Netherlands Washington
DC
and USA
Tilburg University Wesley Yung

Tilburg Statistics Canada
The Netherlands Ottawa
Canada
David B. Wasshausen
Chief, Expenditure and Income Division Stefan Zins
Bureau of Economic Analysis German Institute for Employment Research
USA Nuremberg
Germany
Deanna Weidenhamer
US Census Bureau
k Washington k
DC
USA
k
1
Section 1
Introduction to New Measures/Indicators for the Economy

3
Advances in Business Statistics, Methods and Data Collection:

Introduction
Ger Snijkers 1 , Mojca Bavdaž 2 , Stefan Bender 3,4 , Jacqui Jones 5 , Steve MacFeely 6,7 ,
Joseph W. Sakshaug 8,9 , Katherine J. Thompson 10 , and Arnout van Delden 1
1
Department of Research and Development (Methodology), Statistics Netherlands, Heerlen/the Hague, The Netherlands
2 Academic Unit for Mathematics, Statistics, and Operations Research, School of Economics and Business, University of Ljubljana,
Ljubljana, Slovenia
3
Data Service Center, Deutsche Bundesbank, Frankfurt, Germany
4
School of Social Science, University of Mannheim, Mannheim, Germany
5 Macroeconomic Statistics Division, Australian Bureau of Statistics, Canberra, ACT, Australia
6
Department of Data and Analytics, World Health Organization, Geneva, Switzerland
7
Centre for Policy Studies, University College Cork, Cork, Ireland
8
Statistical Methods Research Department, Institute for Employment Research, Nuremberg, Germany
9 Department of Statistics, Ludwig Maximilian University of Munich, Munich, Germany
10
Associate Directorate for Economic Programs, U.S. Census Bureau, Washington, DC, USA
1.1 The ICES-VI Edited Volume: A New Book on Establishment

Statistics Methodology
In 2021, when ICES-VI took place, almost 30 years had passed since ICES-I, which yielded the
edited volume “Business Survey Methods” (Cox et al. 1995), the first reference book covering this
topic. During these three decades much has changed with regard to establishment statistics,
methods, and data collection. Clearly, it is time for a new volume, discussing these advances.
At the beginning of the 1990s, a group of practitioners working on establishment surveys recog-
nized that the “lack of published methods and communication among researchers was a stumbling
block for progress in solving business surveys’ unique problems” (Cox et al. 1995: xiii) and con-
cluded that “an international conference was needed to (1) provide a forum to describe methods in
current use, (2) present new or improved technologies, and (3) promote international interchange
of ideas” (ibid: xiii). In June 1993, practitioners, researchers, and methodologists from around the
world met in Buffalo, New York (USA), at the International Conference on Establishment Surveys
(ICES) to develop a community based entirely around establishment statistics. Rather than nar-
rowly focus on business surveys, the conference organizers broadened the scope to include all estab-
lishment surveys, whose target populations include businesses (establishments, firms), farms, and
institutions (schools, jails, governments) – basically, any unit other than a household or a person.
This conference created a recurring forum for networking and innovation in an often-overlooked
area of survey research methods and became the first in a series of international conferences.
As this community solidified, and the pace of developments in establishment surveys quickened,
ICES progressed from a seven-year conference cycle to a four-year cycle, with conferences in 2000
(II), 2007 (III), 2012 (IV), and 2016 (V), with the latest conference (ICES-VI) held virtually in
Advances in Business Statistics, Methods and Data Collection. Edited by Ger Snijkers, Mojca Bavdaž, Stefan Bender,
Jacqui Jones, Steve MacFeely, Joseph W. Sakshaug, Katherine J. Thompson, and Arnout van Delden.
© 2023 John Wiley & Sons, Inc. Published 2023 by John Wiley & Sons, Inc.
4 1 Advances in Business Statistics, Methods and Data Collection: Introduction
2021, and ICES-VII being scheduled for 2024. Along the way, the conference title changed, from
“establishment surveys” to “establishment statistics,” reflecting the increased prevalence of other
data sources, such as censuses, registers, big data, and blended data applications in economic
measurements. This series of conferences has become a well-respected international platform
whose participants include practitioners, researchers, and methodologists in government, central
banks, academia, international organizations, and the private sector from around the globe.
The ICES-I edited volume “Business Survey Methods” (Cox et al. 1995) stood for almost
20 years as the only comprehensive overview reference book dedicated to establishment programs.
Featuring chapters drawn from conference presentations, this volume laid the groundwork for
other important references. “Designing and Conducting Business Surveys” (Snijkers et al. 2013) was
published in 2013. This textbook provides detailed guidelines of the entire process of designing
and conducting a business survey, ranging from constructs to be measured using a survey, through
collection and post-collection processing to dissemination of the results. “The Unit Problem
and Other Current Topics in Business Survey Methodology” (Lorenc et al. 2018) was published in
2018. In addition to these monographs, the Journal of Official Statistics published a selection
of 2007 ICES-III papers in a special section (JOS 2010), and subsequently published dedicated
special issues featuring papers from the 2012 ICES-IV (Smith and Phipps 2014) and 2016 ICES-V
(Thompson et al. 2018) conferences.
While the topics covered in that first edited volume remain relevant, much has changed over
the passing decades. As economies change and sciences evolve, methodologists around the world
have worked actively on the development, conduct, and evaluation of modern establishment
statistics programs. Furthermore, emerging data sources and new technologies are used in modern
applications; some of these were not discussed in the monographs noted above, simply because
they did not yet exist or were used on a small scale. These topics include new developments
in establishment surveys (such as nonprobability sampling, sophisticated developments in
web surveys, applications of adaptive and responsive data collection designs), improvements
in statistical process control applications, advanced data visualization methods and software,
widespread use of alternative/secondary data sources like registers and big data along with
improved methodologies and increased production of multi-source statistics, Internet of Things
possibilities (smart farming data, smart industries data), new computer technologies (such as web
scraping, text mining), and new indicators for the economy (like Sustainable Development Goals).
It is time for a new monograph on establishment statistics methodology! The ICES-VI conference
provides an excellent source of material for such a new volume.
ICES-VI was originally planned for June 2020 in New Orleans, but postponed because of the
COVID-19 pandemic, and was held as an online conference from 14–17 June 2021. Close to 425
participants from 32 countries representing over 160 organizations, met online to listen to and
discuss nearly 100 presentations, dealing with a wide variety of topics germane to establishment
statistics. In addition, 17 invited pre-recorded Introductory Overview Lectures (IOL) provided
a comprehensive overview of all key topics in establishment statistics (see Appendix). Book
contributions from conference authors were solicited via an open call to presenters; in some cases,
chapters present consolidated session content. Taken collectively, this edited volume presents
materials and reflects discussions on every aspect of the establishment statistics production
life cycle.
This volume is entirely new; it is neither an update of the Cox et al. book (1995) nor an update
of the Snijkers et al. (2013) textbook. Its scope is broader, containing comprehensive review papers
drawn from IOLs and cutting-edge applications of methods and data collection as applied to
establishment statistics programs, and thoughtful discussions on both previously existing and
new economic measures for economies that are increasingly global. The next section of this
introductory chapter discusses the importance of establishment statistics (Section 1.2), followed by
trends in establishment statistics research based on the ICES conference programs (Section 1.3),
the organization of this volume (Section 1.4), and a conclusion (Section 1.5).
1.2 The Importance of Establishment Statistics
Establishment statistics provide information on businesses, farms, or institutions (Cox and

Chinnappa 1995). These statistics not only explain the structure and evolution of economies
around the world but they also provide many of the input data required by the System of National
Accounts (SNA) and the calculation of gross domestic product (GDP) and contribute to some of
the most important policy questions of our age – for example, how globalization and digitalization
are impacting national economies, societies, and environments.
Establishment statistics come in many “shapes and sizes,” ranging from statistics on a single
industry like agricultural statistics or a single topic like research and development (R&D) statistics
to all-encompassing statistics like structural business statistics (Jones et al. 2013):
• Agricultural statistics help inform business decisions on plants to crop and policy decisions on
subsidies.
• School surveys provide valuable data on the current state of public and private education.
• Government finance statistics provide useful snapshots of spending at the level of local and
federal governments, measure hidden and visible public tax collection, and demonstrate acces-
sibility to common utilities.
• Structural business statistics (SBS) answer key policy questions regarding wealth and employ-
ment creation in different industries and sectors of the business economy and whether these
are growing or declining. They show the costs of doing business in a country, the structure of
earnings, and provide critical information on the specialization of national economies, the prof-
itability of industries, sectors and countries, and changes to productivity and competitiveness.
They offer insights into structure, performance, and dynamics by industry, region, and size.
• Short-term statistics (STS) provide a wide range of sub-annual statistics (typically quarterly or
monthly) to support policy interventions by providing up-to-date statistics on the business econ-
omy. Often published as indices, short-term statistics typically include prices (e.g. consumer,
producer, and residential property inflation), retail sales, manufacturing and services produc-
tion, and income statistics.
• International goods and services trade statistics, provide a wealth of information on the volume
and value of tangible and intangible bilateral trade flows. With increased intensity of cross-border
activities, other statistics have also become relevant:
⚬ Foreign affiliates statistics (FATS), which describe the structure and activity of enterprises
resident in one country but controlled from another – these statistics are of crucial importance
to understand the globalization puzzle.
⚬ Foreign direct investment (FDI) statistics, often collected as part of the Balance of Payments,
supplement the FATS perspective, and reveal the real impact investment when pass-through
or phantom investment and reversable investment are excluded.
⚬ The complexity of international trade and the location of value-added activities have been
explored with the development of global value chains (GVCs). GVCs can be analyzed from
a macro perspective, through the use of global input–output (IO) tables, or from a micro
perspective by examining business function statistics. Linking business registers and trade
statistics (noted above) also contributes to these analyses.
Taken together, these establishment statistics provide a detailed and comprehensive picture of
the economy. They are essential for the compilation of the National Accounts, Input–Output, and
Supply–Use tables. The SNA framework, which is frequently used by policymakers, business com-
munities, and economists, is discussed in Sections 1 and 2 of this book. Advances in measurement
on phenomena like globalization, climate change, e-commerce, and informal sector of the economy
are needed. These topics are discussed in Section 1. While traditional establishment statistics are
produced using surveys and administrative data, resulting in multisource statistics (as discussed in
Sections 2–4, 6, and 7), the production of these new statistics may profit from new data sources and
new technologies, like big data, web scraping, and text mining (Bender and Sakshaug 2022; Hill
et al. 2021), as we will see in Sections 5 and 7.
Whatever data sources used, statistics that target organizations must deal with unique fea-
tures that present entirely different challenges from their household statistics counterparts
(Cox and Chinnappa 1995; Snijkers et al. 2013; Snijkers 2016). Even though establishment
survey methodology borrows from social survey methodology, blindly applying the same survey
methods, collection, and statistical models to establishment statistics can yield disastrous results.
A typical – and longstanding – problem in business statistics is the mapping of unit types that
are found in secondary data sources (like administrative data sources) and the statistical units
that are economically meaningful and comparable over time and across countries, as is shown in
Figure 1.1. The ICES-I edited volume (Cox et al. 1995) devoted a chapter to the construction of
statistical unit types held in a statistical business register (SBR) (Nijhowne 1995), and a chapter
on changes of those unit types over time (Struijs and Willeboordse 1995).
To reduce respondent burden on surveyed establishments or to validate survey collections,
establishment programs often utilize administrative data collected from the same (or nearly the
same) organizations about their operations, such as Value Added Tax (VAT) or tax data. Linkages
Reporting n 1 Enterprise
unit group
Tax unit
1
n
m
Reporting n 1 1 n
Enterprise Legal unit
unit
1
Reporting n 1 Local
unit unit
Reporting Statistical Administrative

unit unit unit
Figure 1.1 The unit problem in establishment statistics: A schematic representation of reporting,
statistical, and administrative unit types. Lines represent relations between the unit types: (1 : n), (n : 1), and
(m : n), with m, n ≥ 1. The exact situation and terminology may differ between countries. Source: Inspired by
Figure 13.1 in Chapter 13 by Cox et al. 2023.
to the corresponding administrative entities are not always straightforward. Governmental

administrative data often contain unit types that are combinations of legal units, e.g. tax units that
may have a many-to-many relationship with statistical units, as is shown in Figure 1.1 (right side;
Cox et al. 2023 [Chapter 13]). In practice, errors may arise when identifying, characterizing, and
delineating statistical units, which are referred to as unit errors in van Delden et al. (2018) i.e.
errors in the relation (and subsequent linkage) between administrative and statistical units. Van
Delden et al. emphasize the importance of assessing the impact of unit errors on the quality of the
output establishment statistics, as they affect accuracy and inference.
A related unit problem exists in surveys where the collected data should match the statistical
unit. Many establishment surveys have differing survey (sample) units and reporting units: “A sur-
vey unit is a business selected from the underlying statistical population of similarly constructed
units (i.e. from the sampling frame). A reporting unit is an entity from which data are collected.”
(Thompson et al. 2015: 774; Haraldsen 2013). Within the same survey, the reporting unit can be
the unit sampled from the frame or another entity entirely and can provide information for itself
or for other units. While survey samples are mostly based on statistical unit types depending on
the sample frame used, sampled units may be split into multiple reporting units (see Figure 1.1,
left side). That typically occurs for large complex businesses that have multiple administration
systems, as is illustrated in Chapter 16 by Snijkers et al. (2023: Figures 16.5 and 16.6). This has
major consequences for the survey design as discussed in Chapter 15 by Haraldsen (2023). Unit
errors in surveys, i.e. failure to report on all subunits or errors on the subunit reporting, can lead
to inadvertent bias in the produced statistics (Snijkers 2016; Zhang 2012).
Household surveys rely on master address files, requiring up-to-date information on whether
a structure is a personal dwelling or an out-of-scope establishment. Establishment surveys
develop a register, merging statistical and administrative data to create complete listings in
real time (Smith and Yung 2023 [Chapter 27]; Machdi et al. 2023 [Chapter 9]). Registers track
changes in survey unit composition as well as additions (births) and deletions (deaths). For
example, a business structure might change because of mergers, acquisitions, or divestitures,
a business could change its economic activity, a farm could grow new crops, or a local school
building could become an administration building. Maintaining these linkages and updates
is paramount, as perhaps the most basic or fundamentally important establishment statistics
are business register and business demographic statistics. These kinds of information are the
backbone of economic statistics, providing the statistical infrastructure upon which almost all
other business statistics rely. Not least, business registers provide the enterprise population
from which other business survey samples are drawn and extrapolated to. Business register
and business demographic statistics provide information on the active enterprise population,
and their churn (births, deaths, life expectancy and survival) as well as identify high-growth
enterprises/sectors and recently formed high-growth enterprises, poetically referred to as gazelles.
They also form the basis for small and medium-size enterprises (SME) and entrepreneurial
statistics. By mapping the legal and geographic structure of enterprises, business registers hold
the key to understanding the operations of multinational enterprises, and by extension global-
ization. By linking business registers with trade statistics, further light is shed on globalization
by also allowing policy makers to understand the role of companies in international goods and
services trade.
As establishments modernize and innovate, establishment statistics must likewise continuously
evolve; and as economies develop and change, establishment statistics and their production life
cycle must change to reflect well the reality. These changes over time are clearly reflected in the
ICES series conference programs.
1.3 ICES Trends

The ICES series of conferences is a rich source of information providing insights in the production
of the abovementioned statistics, and the changes therein. While much of the survey research
methods literature focuses on household surveys and statistics, and social survey methodology has
a long-standing history of more than 100 years (Moser and Kalton 1971, Bethlehem 2009, Groves
2011), we can state that the first milestone in the history of establishment survey methodology
was ICES-I (note that official establishment statistics go back much longer; Jones et al. 2013, Jones
and O’Byrne 2023 [Chapter 8]). The ICES website (hosted by the American Statistical Association,
ASA 1993–2021) has become a rich and irreplaceable repository of papers on many topics related
to establishment statistics.
This section highlights some of the fascinating developments in establishment statistics in the
last three decades. Since the first conference, the goal of the ICES series to present and exchange
information and ideas on current and new methods, data sources, and technologies has not
changed: each conference features (for the time) state-of-the-art methods applied to produce
establishment statistics. However, the scope of the ICES conference programs, the breadth of
the methods and topics covered, and the methods themselves have indeed advanced. In these
developments, there is one common trend: integration. The ICES programs have evolved to
provide holistic views of the entire establishment statistics production life cycle, moving from
production centered stove pipe approaches (or silos) toward integrated statistics programs with
integrated production processes, applying generalized methodologies and alternative data sources,
and involving statistics-user and data-provider perspectives.
At the beginning of the ICES series, data collection in establishment statistics (as produced
by National Statistical Institutes, NSIs) was simple: many NSIs conducted (often mandatory)
mail-out/mail-back sample surveys with paper forms or questionnaires. These forms were charac-
terized by detailed items with lengthy definitions containing technical terms often unfamiliar to
the respondent. These definitions stemmed from government or (in Europe) Eurostat regulations
and were often not in line with the record keeping information available from the surveyed unit,
resulting in high response burden and requiring extensive post-field data cleaning. Economy-wide
programs were essentially nonexistent. Instead, establishment surveys were designed to collect
data for a specific set of statistics (as discussed in Section 1.2 above), for a specific branch of
industry, with surveys designed and executed independently of each other in stove pipes or silos.
Survey-specific solutions were developed, essentially in a vacuum, with the exceptions of the
sample units and definitions of classifications (e.g. size class and economic activity; Jones and
O’Byrne 2023 [Chapter 8]). The focus was on the NSIs internal production process; what happened
within businesses was a black box. The majority of ICES-I papers (ASA 1993; Snijkers 2009)
addressed issues dealing with surveying specific branches of industry, like agriculture, energy,
health care, trade, finance, education, manufacturing, and the construction industry. Topics
revolved around business frames, sampling and business registers, classification systems, data
editing/cleaning, analysis, and estimation. Scant attention was given to general data collection
methodology issues like questionnaire design, data quality, and nonresponse.
By the second ICES, a slow shift in establishment statistics research emerged, beginning to trend
toward more generalized methodology, integrated statistics programs, and data integration. That
said, Linacre (2000: 1) noted that “When you first compare the ICES2 program with the ICES1
program, the first thing that is striking is the amount in common. The overall survey process has
not changed much in 5 years, or indeed 25 years, and in many areas, it is a case of slow progress in
continuing directions.” The trend toward integration and the development of a holistic view has
1.3 ICES Trends 9
accelerated since ICES-II. While the conference series deserves some credit for providing a forum
for collaboration, the drivers for these changes are, as ever, reducing production costs, reducing
response burden (i.e. reporting costs), improving data quality, and – above all – meeting emerg-
ing and increasing user demands for more timely, detailed, and coherent statistics (Linacre 2000;
Edwards 2007; Jones and O’Byrne 2023 [Chapter 8]).
An important integration trend is the need for harmonized and coordinated statistics programs
and integrated processes. At ICES-II, Smith (2000: 23) stated that “the most effective road to
improved statistical quality and minimum response burden lies through greater unification,
harmonization, and integration of statistical programs.” Indeed, many NSIs have been integrating
their statistics programs and production processes with the objective of maximizing production
efficiency, as well as addressing new user demands, both with regard to established economic
statistics and new statistics that adequately measure new economic developments like globaliza-
tion, e-commerce, or the informal economy. Since ICES-II, many of them have presented their
plans and experiences at subsequent ICES conferences.
This push to integrated business statistics production processes certainly accelerated the
development of standardized business survey methods. The ICES-III, IV, V, and VI programs
feature presentations on coordinated sampling designs, integrated business registers, coordinated
data collection, questionnaire design tailored to the business context, electronic questionnaires,
enhanced survey communication tools, generalized data editing and imputation programs, canned
estimation programs, and standardized platforms developed for dissemination of statistics.
Fundamental to this progress is the integration of the establishment perspective into the survey
design, yielding a deeper understanding of the business response process. This was first discussed
at ICES-I and elaborated upon at every subsequent ICES conference. Opening this “black box” has
proved to be instrumental in questionnaire design and pre-testing, in improving survey partici-
pation, implementing electronic data collection instruments, and executing mixed-mode designs,
thus reducing costs and response burden, and improving data quality.
The use of alternative data sources and data integration is another ICES integration trend.
From ICES-I onward, the expanded use of administrative data in establishment statistics has
been discussed. As Linacre mentions at ICES-II (2000: 3): “This area is receiving a lot of
attention by methodologists and good advances are being made in a number of countries.”
Historically, administrative data were used to construct sampling frames and business registers.
However, methodologists and NSIs systematically promoted exploitation of administrative data
to supplement or even replace survey collection, with an on-going objective to reduce response
burden.
By ICES-V, the data integration presentations expanded to include big data and other non-
traditional (organic) data, with an IOL as well as presentations featuring creative small area
estimation applications and investigations of satellite photography data and purchased third-party
registers. ICES-VI featured an even larger percentage of presentations on data integration, with
applications often involving sophisticated machine learning methods, emphasizing the production
of multisource statistics with linked data.
A final integration trend ICES has contributed to, concerns the increased networking and
collaboration between experts from different areas, yielding synergy between these areas. Every
ICES program retains “traditional” topics like respondent contact strategies, questionnaire design
practices, business frames and sampling, weighting, outlier detection, data editing and imputation,
data analysis, estimation, and variance estimation. By blending of standard survey methodology
and survey statistics topics – applied exclusively to establishment data – and emerging methods
and technologies, the ICES conference series has increased what was once a small pool of
specialists to a large and international group of multidisciplinary scientists. After six conferences,
we can conclude that the goal of the founding group (Cox et al. 1995) has been achieved.
These trends in integration at various levels, which are also seen in other international platforms
discussing establishment statistics (see e.g. SJIAOS 2020), will surely continue to be seen at the
next ICES conferences, since the basic challenges remain the same. Over the ICES years, the user
quality criteria for official statistics have been fully developed, as described by Marker (2017; Jones
and O’Byrne 2023 [Chapter 8]). In the future, criteria like timely, relevant, accurate, and coherent
statistics will remain to be critical drivers for innovating production procedures and methods. In
addition, production efficiency considerations, like costs, time, capacity, and response burden, as
well as data quality issues as defined by the total survey error framework (Haraldsen 2013; Snijkers
2016) and other data quality frameworks (for administrative data and big data: van Delden and
Lewis 2023 [Chapter 12], Biemer and Amaya 2021; Amaya et al. 2020) will define the production
conditions (Jones and O’Byrne 2023 [Chapter 8], Bender et al. 2023 [Chapter 22]).
At ICES-I Ryten (1995: 706) concluded, when discussing “Business Surveys in Ten Years from
now,” that “national statistical agencies must adjust today’s structure of surveys, censuses, and
administrative registers as well as today’s capabilities.” We have seen this trend during the ICES
years. However, moving away from siloed statistics production took time, and even today many
NSIs are still organized along traditional statistics outputs. The 2020 COVID pandemic demon-
strated the fluidity of procedures and methods in NSIs under pressure, producing concurrent
measures of generally good quality and yielding new and important statistics at the same time
(see Chapter 11 by Jones et al. 2023). We expect this pandemic to be a tipping point, and that in
the future NSIs will be more flexible in producing more timely and relevant statistics, using and
integrating various data sources, and applying a number of statistical and data collection methods.
(For additional discussions on the impact of the COVID-19 pandemic on official statistics, see the
Statistical Journal of the IAOS special issues on this topic [SJIAOS 2021a]).
The advantage of today, as compared to the times of ICES-I, is that our methods have developed,
more data sources can be used and better exploited, and IT technology has improved, as we have
seen above. Now we need to make sure that all these methods, technologies, and data sources come
together. Integrating new IT technologies in our methods is a next step we need to take (Bavdaž et al.
2020).
Looking to the future we expect surveys to be better integrated and coordinated, using more
sophisticated sample designs and sophisticated estimation methods that benefit from AI (artificial
intelligence) and machine learning technologies. Instead of using questionnaires, to a large extent
the data will be gathered using System-to-System data communication methods (or Electronic Data
Interchange, EDI methods; Buiten et al. 2018) applying for example eXtened Business Reporting
Language (XBRL) protocols or Application Programming Interfaces (API). In the next decade, we
expect businesses themselves to be ready for this technology, having implemented smart industry
technologies (or Industry 4.0; Chakravarti 2021; Haverkort and Zimmermann 2017) using sensor
technology and the Internet of Things yielding an integrated business information chain (Bharosa
et al. 2015), and having developed harmonized data definitions. As Snijkers et al. (2021) concluded
from an exploratory study with precision farming data: “The fruit is not hanging as low as we
thought”: wide-scale adoption of smart industry technologies and data harmonization between
various platforms is inevitable, even if not widely adopted in June 2021. Like with administrative
registers, NSIs should work closely together with other parties like software developers of business
systems to implement EDI systems, making system-to-system data communication possible, and
having the required data stored in the business systems. Also, web scraping and text mining may be
used to collect data directly from the internet. An example of the use of web scraping in the field of
1.4 Organization of This Book 11
establishment statistics is the ADIMA database on multinational enterprises as developed by the

OECD (Ahmed et al. 2019; see also https://www.oecd.org/sdd/its/statistical-insights-the-adima-
database-on-multinational-enterprises.htm). Another example is climate change, where new IT
technologies, new data sources and their integration can help to produce the requested statistics,
although “in the short- and medium term such data have to be extracted and collected from mostly
unstructured data sources.” (Bender and Sakshaug 2022: 15). With all these new IT technologies,
we need to keep in mind that it is the technology that makes new developments possible, but it
is the methodology that makes it work. Ultimately, IT technology is a mere tool to an end, while
methodological assessments are needed to evaluate and improve new data collection methods, new
data sources, and new estimation methods according to cost requirements and quality frameworks.
A final future integration trend is improved collaboration at the international level, among
countries in the developed world but also with developing countries. Again, we can quote Ryten
(1995: 706), who states: “The days of strictly national development of business surveys [read:
establishment statistics] may be over. Instead, we may be entering an era of joint ventures involving
groups of countries organized within supranational entities or motivated by cultural similarities
and trading relations.” International organizations (see Section 2), like the United Nations (UN),
the Organisation for Economic Co-operation and Development (OECD), the World Bank Group,
and the European Union (EU), among others, are trying to establish this, but we still have a long
way to go. Today, we even see some counter developments, considering e.g. the exit of the UK from
the EU. For developing countries, international cooperation is critical to set up intensive programs
in capacity building to achieve good working systems of national statistics (see SJIAOS, 2021b, for
a discussion on training in statistics).
Producing establishment statistics has developed into a multidisciplinary science, closely related
to what nowadays is called Data Science, needing highly educated staff, and it will – or even
should – continue to do so. This book reflects the integration trends discussed above and shows the
multidisciplinary character of this science. It gives the reader a flavor of the richness, complexities,
and variety of facets involved in the production of establishment statistics.
1.4 Organization of This Book

“Advances in Business Statistics, Methods and Data Collection” provides a broad overview of
currently available methods, new developments, and challenges for establishment statistics, with
manuscripts prepared by practitioners and methodologists from across the globe.
This book is divided into seven sections, each section discussing a major theme related to the
production of establishment statistics:
● Section 1: Introduction to New Measures/Indicators for the Economy
● Section 2: Topics in the Production of Official Establishment Statistics and Organizational Frame-
works
● Section 3: Topics in the Use of Administrative Data
● Section 4: Topics in Business Survey Data Collection
● Section 5: Topics in the Use of New Data Sources and New Technologies
● Section 6: Topics in Sampling and Estimation
● Section 7: Topics in Data Integration, Linking and Matching.
The first two sections set the overall framework for this volume: Section 1 discusses topics from
the user perspective – the demand side, whereas Section 2 discusses topics related to the produc-
tion side of statistics – the supply side. Sections 3–5 discuss the three broad data sources used in
Another random document with
no related content on Scribd:
upon them a ray which they will reflect and send back to his own
eye, he will see in the centre of each a spot with the brightness of
burnished silver. He will be tempted to think that this spot is a bubble
of air; but, by immersing the ice in hot water, you can melt the ice all
around the spot,—and when it alone remains, you will see it diminish
and disappear without any trace of air. The spot is a vacuum. Such is
the faithfulness to herself with which Nature operates; thus, in all her
operations, does she submit to her own laws. We know that ice, in
melting, contracts; and here we arrest the contraction, as it were, in
the very act. The water of the flowers cannot fill the space occupied
by the ice which by its fusion has given birth to them; hence the
production of a vacuum, the inseparable companion of each liquid
flower.
The fragment of compact ice whose elements assume such
beautiful crystalline forms is itself a crystal. This was shown by Sir
David Brewster, who employed for the purpose of analysis that
modified form of light which we call polarised light. It is singularly well
adapted to bring out the peculiarities of the main structure of
substances, owing to the coloured figures which it outlines on a
screen after passing through them. All crystals with an axis—such,
for instance, as Iceland spar—yield a series of brilliantly-tinted rings,
traversed by a regularly-formed cross entirely black. As ice produces
the same figures, we are justified in attributing to it the same kind of
crystallization. We must note, however, that we are referring now to
the thick ice formed on our canals and lakes. If we examined the first
film formed on the surface of the water, we should discover in it a
completely irregular crystallization, the ray of polarised light
producing only a mosaic of varied tints, distributed without any order.
But it is easy to explain the way in which this primary crust or film is
produced. Those portions of the fluid mass in contact with the air are
the first to freeze, but each molecule of ice abandons its heat to the
contiguous water, which thereby is slightly raised in temperature, and
the result is a partial congelation. The surface we are examining then
presents a network of fine needles intercrossed in every direction,
and forming a kind of delicate lace, the meshes or intervals of which
are gradually filled up. When the network is transformed into a
continuous sheet, the loss of heat is diminished more and more as
this external crust grows thicker and thicker; but the development of
the ice invariably takes place by means of long interlaced needles,
as the reader may see for himself by breaking off a portion from the
nearest pond (in winter), and examining the sectional surface.
Having said thus much in reference to the crystallization of ice
and snow, we proceed to explain the regelation and moulding of ice.
Some years ago, Faraday astonished the scientific world by a very
curious experiment. Splitting into two parts a piece of ice, he brought
together the parts at the moment that fusion took place on their
surfaces, and they united immediately. How are we to account for
this effect, which can be produced even in hot water?
When the temperature of water rises, the surface molecules first
become liquid, then gaseous; being placed beyond the coercitive
action of the surrounding particles, they are easily set free;
transported, on the contrary, into the centre of the mass, they are
brought absolutely under the influence of this action, which induces a
new solidification,—or, to use the scientific term, a regelation. In this
way it becomes easy to understand how very various forms can be
communicated by simple pressure to a fragment of ice. If the
observer successively places a straight bar in moulds of increasing
curvature, he may easily compel it to assume the shape of a ring or
even of a knot. In each mould, it is true, the ice breaks; but if the
pressure is kept up, the surfaces of the fragments are brought into
contact, and adhere so as to re-establish a condition of continuity. A
snowball may thus be converted into a sphere of ice, and the sphere,
by constant pressure, into a cup or a statue.
Professor Tyndall refers to a remarkable instance of regelation
which he observed one day in early spring. A layer of snow, not quite
two inches thick, had fallen on the glass roof of a small conservatory,
and the internal air, warming the panes, had melted the snow so far
as it was in immediate contact with them. The entire layer had
slipped down the pane, and projected beyond the edge of the roof,
without falling, and had bent and curved as required, just like a
flexible body.
MOULDING ICE.
The snow-fields which overspread the upper part of every glacier,

whether in the Arctic Regions or elsewhere, are composed of
crystallized snow, whose fragile, delicate, and fairy-like architecture
endures so long as it remains dry, but undergoes a great
transformation when the sun, melting the upper stratum, allows the
water to interpenetrate its substance. The fluid, congealing anew
during the night, transforms the snow into the condition technically
known as névé; a term given by the Swiss physicists to a granular
mass composed of small rounded icicles, disaggregated, but more
adhesive than snow-flakes, and of a density intermediate between
that of snow and that of ice. Under the pressure of new layers, and
as a result of infiltrations of water, the névé unites, and solders into
ice of constantly increasing compactness.
But glacier-ice presents some other curious peculiarities. Every
abundant snow-fall on the summit of the mountains forms a layer
easily distinguishable from preceding layers—which, in most cases,
have already passed into the névé condition. This stratification
becomes more apparent when the whiteness of the surface has
been sullied by dirt or dust wafted on “the wings of the wind.” It is
perceptible also in ice; but here we must not confound it with another
phenomenon of which the cause is different, the veined structure.
In places where glaciers have been accidentally cut down in an
almost vertical direction, the section is found to exhibit a series of
parallel veins, formed by a beautiful and very transparent azure ice
in the midst of the general mass, which is of a whitish colour, and
slightly opaque.
In different glaciers, and in different parts of the same glacier,
these blue veins will vary in number and intensity of colouring. They
are specially beautiful in crevasses of recent formation, and on the
sides of channels excavated in the ice by tiny rills resulting from
superficial fusion. Not a few glaciers exhibit this remarkable veined
structure throughout their entire extent. When a vertical cutting
exposes the delicate azure network to atmospheric influences, the
softer ice melts prior to the fusion of the blue ice which then remains
in their detached leaflets. On examining these attentively, we cannot
fail to remark the absence, or, at all events, the extreme rarity, of air-
bubbles, though they are so plentiful in the coarser ice.
Professor Tyndall’s explanation of this phenomenon is as
interesting as it is ingenious. While on a visit of inspection to the
slate-quarries of Wales, he had occasion to study the cleavage of
the rocks which compose them; in other words, their faculty of
dividing naturally, a property inherent in all crystals. The schistous
slate separates easily into sheets, and in traversing different quarries
one sees that all the planes of cleavage are parallel in each. From
this circumstance our men of science were at first induced to look
upon slates as the products of the stratification of different deposits.
Such an explanation, however, could not be accepted by Tyndall,
when he observed that the minute fossils embedded in them were
constantly misshapen and flattened in the direction of the plane of
cleavage, because the great modification they had undergone could
not have taken place in superimposed strata at the bottom of the
primeval sea. He concluded that these schists, therefore, must have
been subjected to a considerable pressure; and further, that this
pressure must have been exercised at right angles with the plane of
separation of the different layers.
A long series of experiments proved that many bodies, when
forcibly compressed, exhibit in their structure a very distinctly marked
lamination, and frequently veins of very great beauty.
He carefully examined iron which had passed under the steam-
hammer, or through the rolling-mill; clay and wax were subjected to
the hydraulic press. In all cases he detected signs of cleavage; and
hence we are justified in the inference that the phenomenon is
invariably produced by pressure in all bodies of irregular internal
structure. Such is the result with glacier-ice, from whose mass the
air-bubbles introduced by the snow are gradually expelled. At first of
brilliant whiteness, it assumes, in the parallel layers corresponding to
the planes of cleavage, those beautiful azure tints which characterize
the veined structure. So little has it to do with stratification, that in
places where this is apparent it has given rise to a series of
horizontal lines, while the parallel veinings, in the same masses of
ice, are all inclined at an angle of about 60°.
The tendency to cleavage in compact ice would seem to explain
the regular form of those fragments or detached pieces with which
some parts of the glaciers are covered. Usually they occur as cubes,
or as rectangular parallelopipeds. The Alpine mountaineers name
them séracs,—in allusion to their resemblance to certain cheeses
which bear this name, and which are manufactured in rectangular
boxes. They have been found in many parts of a really colossal size,
measuring fifty feet in length, breadth, and depth, and as regular in
shape as if they had been hewn with a chisel.
There are many interesting points connected with the formation
and constitution of glaciers which we should gladly discuss, but we
are confined by our limits to remarks of a general character, and we
must now pass on to speak of the phenomena attendant upon their
motion. No doubt, the traveller who for the first time comes in sight of
one of these huge ice-rivers, and sees the mighty mass apparently
rooted to its valley-bed, solid, unchangeable, adamantine, finds it
hard to believe that it moves onward with a certain and an unresting,
though a gradual progress. It looks like a noble river, suddenly
petrified by some overwhelming force: congealed, as it flowed, in a
moment, by some irresistible spell! Such, indeed, is the conception
of the poet:—
“Ye ice-falls! ye that from the mountain’s brow
Adown enormous ravines slope amain....
Torrents, methinks, that heard a mighty voice,
And stopped at once amid their maddest plunge!
Motionless torrents! silent cataracts!”
And this conception is justified by the aspect of the glacier. Thus,

of the Glacier du Géant, Professor Tyndall says:—“It stretches
smoothly for a long distance, then becomes disturbed, and then
changes to a great frozen cascade, down which the ice appears to
tumble in wild confusion. Above the cascade you see an expanse of
shining snow, occupying an area of some square miles.” But we shall
see that here, as in the world of man, appearances are deceitful, and
that the glacier well deserves to be called an ice-river, in allusion to
its regular and continuous motion.
Between the snow-fall in the higher regions of the globe, and the
quantity of snow which every summer disappears through
liquefaction, the difference is very considerable. The supply, so to
speak, exceeds the demand, and a residuum is annually left. It is
only below the perpetual snow-line that the snow created and
accumulated in winter is wholly melted in the warm season. And,
therefore, if for any considerable period the excess upon any
particular mountain continued to accumulate, immense masses of
ice would gradually rise to the extreme height in the atmosphere
affected by aqueous phenomena.
Rendu, the Roman Catholic prelate, who first led the way to the
discovery of the true nature of glaciers, says, very justly,—“The
economy of the world would be soon destroyed, if at certain points
accumulations of matter prevailed. The centre of gravity of the globe
would be insensibly displaced, and the admirable regularity of its
movements would be succeeded by disorder and perturbation. If the
Poles did not send back to the Equatorial seas the waters which,
reduced into vapour, issue daily from these burning regions, to be
converted into ice in the Arctic and Antarctic Zones, ocean would be
drained dry, and life would cease, as well as water, to circulate
throughout our world. The Creator, however, in order to ensure the
permanence of His almighty work, has called into existence the vast
and powerful law of circulation, and this law the careful observer
sees reproduced in all the economy of Nature. The water circulates
from the ocean into the air, from the air it spreads over the earth, and
from the earth it passes into the seas. The rivers return from whence
they came, in order that they may issue forth anew; the air circulates
around the globe, and, as it were, upon itself, passing and repassing
successively at all the altitudes of the atmospheric column. The
elements of every organic substance circulates in changing from the
solid to the liquid or aëriform state, and in returning from the latter to
the state of solidity or organization. It is not improbable that the
universal agent which we designate under the name of fire, light,
electricity, and magnetism, has probably also a circle of circulation
as extensive as the universe. Should its movements ever be known
to us more than they now are, it is probable that they would afford
the solution of a host of problems which still defy the intellect of man.
Circulation is the law of life, the method of action employed by
Providence in the administration of the universe. In the insect, as in
the plant, as in the human body, we find a circulation, or rather
several circulations,—blood, humours, elements, fire, all which enter
into the composition of the individual.”
However fanciful may be some of the amiable prelate’s
speculations, it is certain that the glaciers obey this law of circulation.
The snow-accumulations in the upper regions are to some extent
reduced by the descent of the avalanches,—that is, of masses of
snow and ice which detach themselves from the mountain-sides and
dash headlong into the valleys below, where they are rapidly melted
by the warmer atmosphere. But this would, in itself, be wholly
insufficient. Another movement, at once more efficacious and more
regular, is necessary; a movement which embraces the entire
system of the ice-masses, and which carries the glaciers below the
perpetual snow-line, so that every year they may give up a portion of
their terminal extremities. The discovery of this general progression
is one of the most fertile with which, of late years, the physics of the
globe have been enriched.
Professor Tyndall rightly observes that there are numerous
obvious indications of the existence of glacier-motion, though it is too
slow to catch the eye at once. The crevasses change within certain
limits from year to year, and sometimes from month to month; and
this could not be if the ice did not move. Rocks and stones also are
observed, which have been plainly torn from the mountain-sides.
Blocks seen to fall from particular points are afterwards noticed lower
down. On the moraines rocks are found of a totally different
mineralogical character from those composing the mountains right
and left; and in all such cases strata of the same character are found
bordering the glacier higher up. Hence the conclusion that the
foreign boulders have been floated down by the ice. Further, the
ends or “snouts” of many glaciers act like ploughshares on the land
in front of them, overturning with irresistible energy the huts and
châlets that lie in their path. Facts like these have been long known
to the inhabitants of the High Alps, who were thus made acquainted
in a vague and general way with the motion of the glaciers. But
Science cannot deal with generalities: it requires precise and
accurate information; and this information, so far as the progression
of the glaciers is concerned, has been obtained through the patient
labours of Rendu, Charpentier, Agassiz, Desor, Vogt, Professor
Forbes, Bravais, Charles Martins, Hopkins, Professor Tyndall,
Colomb, John Ball, and Schlagintweit. Their experiments and
observations have established the truth of certain immutable
principles, and proved the existence of a general law of movement.
The accumulation of the débris hurled headlong by the mountains
forms on the glacier-surface long lines of stone and earth, which are
called moraines; these diverge in certain directions, according to the
circumstances we now come to explain.
The landslips which occur on the banks or edges of the glacier
give rise to the lateral moraines, which are enlarged and extended
daily by the twofold effect of the fall of stones and débris, and the
progressive movement which carries them along with the whole
mass of ice. Towards the centre of the great glaciers, in almost every
case, is found a medial moraine; the result of the encounter of the
lateral moraines of two glaciers which have united into one. These
superficial moraines participating in the movement of the glacier,
each of their blocks eventually rolls to the foot of the terminal
precipice, and thus a frontal moraine is formed on the very soil of the
valley, like an embankment raised to prohibit the further advance of
the ice. And, lastly, the bed of sand, gravel, pebbles, and detritus
which is found beneath the glacier, and over which it glides, is called
the profound moraine.
The furrows wrought by this last-named stratum on the bottom of
the glacier-channels show the wonderful force of friction which the
glacier exercises during its descent. The depths of these furrows
depends entirely on the hardness of the débris carried down by the
glacier, and the nature of the rocks submitted to the friction. The
polish assumed by these rocks when they are sufficiently solid to
resist the thunderous march of the glacier, indicates the enormous
pressure which it exercises on the slopes of the valley through which
it forces its way. This effort, bearing principally on the side of the
rocks turned in the direction of their crests, impresses upon them a
peculiar rounded form, so like the appearance of a flock of sheep
(moutons) that De Saussure gave them the name of roches
moutonnées.
Connected with the scientific evidence of the progressive
movement of glaciers, a glacier in the Bernese Oberland will for ever
be memorable. Two branch glaciers, the Lauteraar and the
Finsteraar, unite at a promontory called the Abschwung to form the
trunk-glacier of the Unteraar, which carries a great medial moraine
along its colossal back.
Here in 1827, an “intrepid and enthusiastic” Swiss professor,
Hugi, of Solothurm (or Soleure), erected a small cabin of stones for
the purpose of observations upon the glacier. The hut moved, and he
took steps to measure its motion. In three years, 1827 to 1830, it
moved 330 feet downwards. In 1836 it had descended 2354 feet;
and in 1841, it had accomplished a journey of 4712 feet. [This was at
the rate of about 336 feet a year.]
In 1840, M. Agassiz, with some scientific friends, Messrs. Desor,
Vogt, and Nicolieb, established themselves under a great
overhanging slab of rock on the same moraine, and by means of
side walls, and other appliances, constructed a rough abode which,
because some of these men of science came from Neufchâtel, they
named the “Hôtel des Neuchâtelois.”
In two years after its erection, Agassiz discovered that it had
moved downwards no less a distance than 486 feet.
These and some similar measurements brought to light a very
important fact. The reader will observe that the middle numbers,
corresponding to the central portion of the glacier, are the largest:
hence it was obvious that the centre of a glacier, like that of a river,
moves more rapidly than the sides.
Owing to the greater central motion of a glacier, its crevasses
invariably assume a curved outline, of which the convexity advances
towards the bottom of the valley.
It has also been ascertained that the superficial part of a glacier
moves more rapidly than its base.
Again: Tyndall and Hirst, by employing instruments of great
precision, have demonstrated that the maximum of motion is not to
be found exactly in the centre, but that, according to the windings of
the valley through which the glacier flows, it moves sometimes to the
right of the centre, and sometimes to the left. Now, the progression
of a river exhibits all the characters we have just enumerated, and
the truth foreshadowed by Rendu has been confirmed in every
detail. The glacier is a “river of ice.”
The reader will naturally ask, How can a substance of such
apparent rigidity as ice obey, as it does obey, the same laws which
regulate the movement of fluids? I can understand, he may say, how
water flows in such and such a manner: it is a liquid, and its
molecules are deficient in the property of cohesion; but that so solid,
and firm, and unimpressible a substance as ice should be capable of
motion seems impossible. I can understand very easily that a mass
of ice, when loosened or detached from its resting-place, will glide
downwards until arrested by some adequate obstacle; but this is not
the kind of motion you are describing. According to your
explanations, every constituent portion of the glacier moves, and the
central faster than the lateral, and the surface faster than the base.
These objections were advanced by men of science when the
motion of glaciers was first put forward as a theory; and the answer
given by Scheuchzer was, that a glacier might be compared, in the
summer season, to a sponge saturated with water, which, when
afterwards congealed by the cold temperature of autumn and winter,
expanded, and produced a dilatation of the mass in every direction.
Then, as it could not recede, as it could not reascend its valley-
slope, the augmentation of size would necessarily take place in its
lower portion.
It is unnecessary for us to explain why this answer was
unsatisfactory. Subsequent observations, however, proved its
impossibility, and Professor Forbes then put forward his ideas of the
viscous character of ice. But these, too, did not meet the conditions
of the phenomenon; and the view now adopted is that of Professor
Tyndall, who has shown that it is the result of the regelation we have
already described.
Professor Forbes enunciated his theory in words to the following
effect: “A glacier is an imperfect fluid or viscous body, which is urged
down slopes of certain inclination by the natural pressure of its
parts.” But we know the exceeding brittleness of ice, and how is
viscosity compatible with brittleness? We know, too, that crevasses
and fissures will suddenly form on a glacier, like the cracks on a
pane of glass. But if ice were viscous, and could expand, dilate, or
stretch as viscous substances do, these crevasses would be
impossible. They would gradually close up, like an indent in a mass
of jelly. And yet it cannot be denied that a glacier does move like a
viscous body; the centre flowing past the sides, the top flowing over
the bottom, while the motion through a curved valley corresponds to
fluid motion. How are we to reconcile these apparently conflicting
circumstances?
By Professor Tyndall’s regelation theory, which is founded on a
fact already mentioned; namely, that when two pieces of thawing ice
are brought in contact, they freeze together.
This fact, and its application irrespective of the cause of
regelation, may be thus illustrated: “Saw two slabs from a block of
ice, and bring their flat surfaces into contact; they immediately freeze
together. Two plates of ice, laid one upon the other, with flannel
round them overnight, are sometimes so firmly frozen in the morning
that they will rather break elsewhere than along their surface of
junction. If you enter one of the dripping ice-caves of Switzerland,
you have only to press for a moment a slab of ice against the roof of
the cave to cause it to freeze there and stick to the roof.
“Place a number of fragments of ice in a basin of water, and
cause them to touch each other; they freeze together where they
touch. You can form a chain of such fragments; and then, by taking
hold of one end of the chain, you can draw the whole series after it.
Chains of icebergs are sometimes formed in this way in the Arctic
seas.”
From these observations we deduce the following result:—Snow
consists of small particles of ice. Now, if by pressure we squeeze out
the air entangled in thawing snow, and bring the little ice-granules
into close contact, they may be expected, as they do, to freeze
together; and should the expulsion of the air be complete, the
squeezed snow will assume the appearance of compact ice.
It is in this way that the consolidation of the snows takes place in
the Arctic as in the higher Alpine regions. The deeper layers of the
névé are converted into more or less perfect ice by the pressure of
the superjacent layers; and further, they are made to assume the
shape of the valley which they fill, by the slow and continuous
pressure of its sides.
In glaciers, as Professor Tyndall points out, we have ample
illustrations of rude fracture and regelation; as, for example, in the
opening and closing of crevasses. The glacier is broken on the
cascades, and mended at their bases. When two branch glaciers lay
their sides together, the regelation is so firm that they begin
immediately to flow in the trunk glacier as in a single stream. The
medial moraine gives no indication by its slowness of motion that it is
derived from the sluggish ice of the sides of the branch glaciers.
We may sum up the regelation theory in few words. The ice of
glaciers changes its form and retains its continuity under pressure
which keeps its particles together. But when subjected to tension,
sooner than stretch, it breaks, and behaves no longer as a viscous
body.
These are Professor Tyndall’s words, and the fact which they
embody it would be difficult to set forth more clearly or more
concisely.
A POLAR GLACIER.
Having said thus much of the structure, causes, characteristics,
and movement of glaciers, we proceed to consider some of the more
remarkable of those which are situated in the Arctic World.
The glaciers of the Polar Regions do not differ in structure or
mode of formation from those of other countries. Yet they possess
some peculiar features, and to a superficial observer might seem
independent of the physical laws we have attempted to explain. That
this is not the case has been shown by Charles Martins, who
carefully studied the glaciers of Spitzbergen on the occasion of the
exploring voyage of the Recherche to that island, and has
demonstrated that their differences are but a particular case of the
general phenomenon.
As special characters he points out, first, the rarity of needles and
prisms of ice, which he attributes to the slight inclination and the
uniformity of the slopes, as well as to the diminution of the solar
heat, which, even in the long summer days, does not melt the
surface. There are no rills or streams capable of hollowing out
crevasses and moulding protuberances or projections. But
transversal crevasses produced by the movement of the glaciers are
numerous, and these are often very wide and very deep.
GLACIER, ENGLISH BAY, SPITZBERGEN.

In the terminal escarpment, which melts in proportion as it
plunges into the sea, immense caverns are sometimes seen;
caverns so immense that the azure-gleaming grottoes of the
Arveiron and Grindelwald, so much admired by European travellers,
are but miniatures. “One day,” says Charles Martins, “after having
ascertained the temperature of the sea off the great glacier of Bell
Sound, I proposed to the sailors who accompanied me to carry our
boat into its cavern. I explained to them the risk we should incur,
being unwilling to attempt anything without their consent. When our
boat had crossed the threshold, we found ourselves in an immense
Gothic cathedral; long conical-pointed cylinders of ice descended
from the roof; the recesses seemed so many chapels opening out of
the principal nave; broad fissures divided the walls, and the open
intervals, like arches, sprang towards the summits; azure gleams
played over the icy surface, and were reflected in the water. The
sailors, like myself, were dumb with admiration. But a too prolonged
contemplation would have been dangerous; we soon regained the
narrow opening through which we had penetrated into this winter
temple, and, returning on board our vessel, preserved a discreet
silence respecting an escapade which might have been justly
blamed. In the evening, we saw from the shore our cathedral of the
morning slowly bend forwards, detach itself from the parent glacier,
crash into the waves, and reappear in a thousand blocks and
fragments of ice, which the retiring tide carried slowly out to sea.”
The Spitzbergen glaciers do not exhibit those numerous
moraines which are observed on the majority of those of
Switzerland.
The mountains, not being very lofty, are buried, as it were, under
their burden of glaciers, instead of preponderating over them, and
seem with difficulty to lift their peaks out of the mass of ice and snow
surrounding them. Consequently, there are no considerable landslips
or falls of earth and stone, which, accumulating along the borders of
the glaciers, might form moraines. Martins is of opinion that the
Spitzbergen glaciers correspond to the upper part of the glaciers of
Switzerland; to so much, that is to say, as lies above the perpetual
snow-line.
Now, he says, the higher we ascend on an Alpine glacier, the
more do the lateral and medial moraines diminish in width and form,
until they taper away and finally disappear under the high névés of
the amphitheatres from which the glacier issues, just as the
mountain torrents often take their rise in one or in several lakes
terraced one above the other.
GLACIER, BELL SOUND, SPITZBERGEN.

For all these reasons, he adds, the medial and lateral moraines
are scarcely conspicuous on the glaciers of Spitzbergen; a number
of stones and boulders may be seen along their sides, and
sometimes in their centre, but the ice is never hidden, as in the Alps,
under the mass of débris accumulated upon it. As for the terminal
moraines, they must be sought at the bottom of the sea, since the
terminal escarpment nearly always overhangs it. Hence, the blocks
of stone fall simultaneously with the blocks of ice, and form a
submarine frontal moraine, of which the two extremities are
occasionally visible upon the shore.
In a previous chapter we have alluded to the manner in which
icebergs are formed by the detachment from the seaward extremity
of the glacier of huge masses of ice, which the current carries out
into the open sea. To the description already given, we may here add
that which Charles Martins furnishes in his valuable and interesting
record of persevering scientific enterprise, “Du Spitzberg au Sahara”:
—In Spitzbergen, he says, the glacier, after a traject of more or less
considerable duration, reaches the sea. If the shore be rectilineal, it
advances no further; but, in the recess of a bay, where the shore is
curved, it continues its progression, supporting its bulk on the sides
of the bay, and advancing above the water, which it overhangs. This
is easily understood. In summer the sea-water at the bottom of the
bays is always at a temperature a little above 32°; on coming in
contact with this comparatively warm water the glacier melts, and, at
low tide, an interval is perceptible between the ice and the surface of
the water. The glacier being no longer supported, partially crumbles
and gives way; immense blocks detach themselves, fall into the sea,
disappear beneath the water, reappear revolving on their own axes,
and oscillate for a few moments until they have taken up their
position of equilibrium. The blocks thus detached from the floating
masses, of all sizes and shapes, are called icebergs.
STEAMER “CHARGING” AN ICEBERG, UPERNAVIK, GREENLAND.

Our traveller records that twice a day, in Magdalena Bay and Bell
Sound, he was an eye-witness of this partial ruin of the extremity of
the glaciers. Their fall was accompanied by a noise like that of
thunder; the swollen sea rushed upon the shore in a succession of
gigantic waves; the gulf was covered with icebergs, which, caught in
the swirl and eddy, issued out of the bay, like immense fleets, to gain
the sea beyond, or were stranded here and there at points where the
water was shallow. The icebergs seen by M. Martins were not,
however, of any surprising magnitude; he estimates their average
height at thirteen to sixteen feet. We have seen that those of Baffin
Bay are tenfold more considerable and imposing; but then, in that
bay the temperature of the sea is below 32°; the glacier does not
melt when it enters the water; it sinks to the bottom of the sea; and
the portions detached from it are all of greater height than even the
submerged part of the icebergs which drift to and fro in the bays and
gulfs of Spitzbergen.
We may follow up this description with some observations by

Lieutenant Bellot, the chivalrous young Frenchman who perished in
one of the expeditions despatched in search of Sir John Franklin and
his companions. He is speaking of the masses of ice his ship
encountered soon after doubling Cape Farewell, the south point of
Greenland, and he remarks, that as Baffin Bay narrows towards the
south, the icebergs, first set in motion higher up the bay by the
northern gales, necessarily tend to accumulate in the gorge thus
formed, and so to impede and block up Davis Strait, even when the
higher waters are quite free. It is only through a series of alternate
movements of advance and recession that the bergs finally pass
beyond the barrier, and float out into the Atlantic, to undergo a slow
process of dissolution.
The mobility of the bergs, though necessary to navigation, forms
at the same time its peculiar danger, since a vessel is often placed
between the shore and the colossal masses driven forward by the
wind, or between these and the solid ice which as yet has not broken
up. It is useless to dwell upon the immense force possessed by
masses which are frequently several square leagues in extent, and
which, once in movement, cannot be stayed by any human
resistance. A sailing-vessel finds herself placed in conditions all the
more unfavourable, because the winds blow from the very direction
which she is bound to take in order to open up a way through the
floes. Now, if the gale is violent, it is perilous indeed to push forward
in the midst of a labyrinth of bergs, which form so many floating
rocks; if a calm prevails, a ship can move forward only by laborious
hauling or towed by the boats. The application of the screw-propeller
to steam-ships has given to them a great superiority, because they
are not liable to any accident to paddle-wheels, exposed as such
must be to collision with the floating ice. It is recorded that, on one
occasion, a screw-steamer, near Upernavik, on the coast of
Greenland, actually charged an iceberg, and drove right through it,
as a railway-engine might crash through a fence or hurdle. Of
course, the berg was of no great elevation; but its solid mass yielded
to the immense force of the steamship, and split into large
fragments.
In the convulsions caused by furious tempests, which are far from
being so rare within the Arctic Circle as is popularly supposed, the
shape of the bergs becomes very irregular, and the configuration of
the ice-fields is constantly undergoing modification. Hence it often
happens that the voyager sees before him an open basin of water of
greater or less extent, from which he is separated only by a narrow
strip of ice. In such a case he endeavours to effect an opening,
either by driving his ship at full speed against the weakest part of the
ice, or with the help of immense saws, twenty feet in length, which
are worked with a rope and pulley placed at the top of a triangle
formed of long poles; or, finally, by exploding a mine. When the ice is
not very solid, the ship is forced into the opening, against the sides
of which it acts like a wedge. It will sometimes occur, in the course of
the operation, that the ice-fields, set in motion by the wind or the
currents, close in together, after having treacherously separated for a
moment, and the vessel is then subjected to a dangerous pressure.
Unhappy the mariner who does not foresee or sufficiently note the
warning signs of this accident, which is almost always accompanied
by fatal consequences. The ice, which nothing can check, passing
underneath the ship, capsizes it,—or, if it resists, crushes it.
We have alluded to the colossal bergs of Baffin Bay. These are

thrown off from the northern glaciers, and particularly from the
enormous ice-river named after Humboldt, which cumbers the
declivities of the Greenland Alps, beyond the 79th parallel. It has
been a frequent source of surprise to navigators that these mighty
masses should float in a contrary direction to that of the ice-fields
which descend with the Polar current towards the Atlantic. They
reascend with such rapidity that they shatter the so-called “ice-foot,”
or belt of ice, still adhering to the shore. Captain Maury has collected
numerous observations on this important subject, and he quotes the
case of a ship which was being laboriously hauled against the
current, when an enormous floating mountain coming up from the
south steered against it, but fortunately did not come into collision
with it, and forging ahead, very quickly disappeared. How is such an
incident to be explained? By the existence of a submarine counter-
current, acting on the lower extremity of the submerged portion of
the berg, which, as we have stated, is always seven or eight times
larger than the bulk above the surface of the waves.
Our whalers, in their hazardous expeditions, often derive
assistance from these moving islands. They seek shelter under their
lee when sudden storms arise; for the huge bergs are scarcely
affected by the most violent gales. They find their shelter valuable
also during certain operations of the fishery for which rest and quiet
are necessary. Yet it is not absolutely exempt from danger. The
seeming friend may prove to be a concealed foe. The iceberg may
collapse, or be capsized; or formidable fragments, loosened from
their sides or summits, may topple headlong and threaten to
overwhelm the ship beneath: but as on these and other accidents we
have already dwelt at length, we refrain from wearying our readers
with a twice-told tale. The repetition in which, to some extent, we
have indulged, was needful, in order to show the reader in what way

Advances in Business Statistics Methods and Data Collection Ger Snijkers Full Chapter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Advances in Business Statistics Methods and Data Collection Ger Snijkers Full Chapter

Uploaded by

Copyright:

Available Formats

Advances in Business Statistics,

Methods and Data Collection Ger

Arnout van Delden

Limit of Liability/Disclaimer of Warranty

Library of Congress Cataloging-in-Publication Data Applied for

Hardback ISBN: 9781119672302

Cover Design: Wiley

Set in 9.5/12.5pt STIXTwoText by Straive, Chennai, India

List of Contributors xxix

Section 1 Introduction to New Measures/Indicators for the Economy 1

1 Advances in Business Statistics, Methods and Data Collection:

2 GDP and the SNA: Past and Present 23

2.2.5 Global Proliferation of GDP 27

3 GDP and the SNA: Future Challenges 43

4 Bridging the Gap Between Business and Macroeconomic Statistics:

5 Measuring Investment in Intangible Assets 79

6 Measuring the US Digital Economy 105

7 Establishment Based Informal Sector Statistics: An Endeavor of

Section 2 Topics in the Production of Ofﬁcial Establishment Statistics

8 Statistical Producers Challenges and Help 147

8.15 Developing a Harmonized Content Model 165

9 The Development and Maintenance of Statistical Business Registers as

9.7.3 System Development 187

10 Managing Response Burden for Ofﬁcial Statistics Business

11 Producing Ofﬁcial Statistics During the COVID-19 Pandemic 225

Section 3 Topics in the Use of Administrative Data 265

12 Methodology for the Use of Administrative Data in Business Statistics 267

13 Developing Statistical Frameworks for Administrative Data and Integrating It

13.3.2 Developing Statistical Pipelines for Processing 300

14 The Evolution of Integrating Administrative Data in Business Statistics in

14.5 The Use of VAT Data in Business Statistics 330

Section 4 Topics in Business Survey Data Collection 335

15 What Computerized Business Questionnaires and Questionnaire

16 Tailoring the Design of a New Combined Business Survey: Process, Methods,

16.5.1.3 Post-field Phase of the First Wave of the Survey 373

17 Advances in Question(naire) Development, Pretesting, and Evaluation 387

18 Using Paradata in Electronic Business Survey Questionnaires 413

19 Recent Findings from Experiments in Establishment Surveys 437

19.5.3 Results 461

20 Web Portals for Business Data Collection 469

21 A Creative Approach to Promoting Survey Response 501

21.4.5 Additional Meetings 509

Section 5 Topics in the Use of New Data Sources and New

22 Statistical Data Production in a Digitized Age: The Need to Establish

22.4.4.4 RDSC Contract Generator 532

23 Machine Learning in German Ofﬁcial Statistics 537

24 Six Years of Machine Learning in the Bureau of Labor Statistics 561

25 Using Machine Learning to Classify Products for the Commodity Flow

25.1.2.1 Nonrespose 573

Section 6 Topics in Sampling and Estimation 613

27 Introduction to Sampling and Estimation for Business Surveys 615

28 Sample Coordination Methods and Systems for Establishment Surveys 637

29 Variance Estimation for Probability and Nonprobability Establishment

30 Bayesian Methods Applied to Small Area Estimation for Establishment

31.4 Multiple Imputation Variance Estimation 711

32 Minimizing Revisions for a Monthly Economic Indicator 727

32.5.3 Imputation Discussion 750

Section 7 Topics in Data Integration, Linking and Matching 755

33 Record Linkage for Establishments: Background, Challenges, and an

34 Methods for Estimating the Quality of Multisource Statistics 781

34.2.1.2 Correcting for Representation Errors 785

35 Adopting Previously Reported Data into the 2022 Census of Agriculture:

36 Integrating Alternative and Administrative Data into the Monthly Business