The Dutch Virtual Census of 2001 Analysis and Methodology Editors Eric Schulte Nordholt Marijke Hartgers Rita Gircour Publisher Statistics Netherlands Prinses Beatrixlaan 428 2273 XZ Voorburg Printed by Statistics Netherlands – Facility Services Cover design WAT ontwerpers, Utrecht Information E-mail: [email protected] Where to order E-mail: [email protected] Internet www.cbs.nl © Statistics Netherlands, Voorburg/Heerlen, 2004. Quotation of source is compulsory. Reproduction is permitted for own or internal use. Prices do not include postage and administration costs. Price: € 48.50 Key figure: B-57 ISBN: 903572469 0 Product code: 6008904010 Explanation of figures . = data not available * = provisional figure x = publication prohibited (confidential figure) - = nil or less than half of unit concerned 0 (0,0) = less than half of unit concerned - = (between two figures) inclusive blank = not applicable 2002–2003 = to 2003 inclusive 2002/2003 = of 2002 up to and including 2003 2002/’03 = crop year, financial year, school year etc. beginning in 2002 and ending in 2003 2000/’01–2002/’03 = financial year etc., beginning in 2000/’01 and ending in 2002/’03 Due to rounding, some totals may not correspond with the sum of the separate figures. Table of contents Preface . 7 1. Introduction to the Dutch Virtual Census of 2001. 9 2. Families at work . 23 3. Working people: what they do, who they are and where they live . 41 4. Young people and their economic activities. 59 5. Active ageing: the senior labour force from work to retirement. 73 6. Foreigners at work . 97 7. The Dutch Virtual Census of 2001 compared to previous censuses. 119 8. Large and medium-sized cities compared Some results of the Urban Audit II. 143 9. Cities in relation to their peri-urban areas Some results of the Urban Audit II. 165 10. Cities under the microscope Some results of the Urban Audit II. 183 11. The Netherlands: commuter country. 203 12. The results of the 2001 Census in the Netherlands, the United Kingdom and some other European countries . 225 13. The use of administrative registers and sample surveys in the Dutch Census of 2001 . 243 14. The method of repeated weighting in the 2001 Census . 261 The Dutch Virtual Census of 2001 5 Preface Under the aegis of the United Nations, most countries in the world conduct national population and housing censuses once every ten years. Some 30 European countries, including the Netherlands, produced identical census tables on 2001. Eventually such census tables will yield a comprehensive socio-demographic and socio-economic view of Europe. Economic activity is a central issue in this book. The authors, most of whom work at Statistics Netherlands 1), discuss the economic activities of families, differences according to family size and composition, individual activity status by region, age, education level and branch of economic activity. There are separate chapters on the economic activities of young people and people of retirement age. The economic activities, levels of education and occupation of foreigners from various countries of origin are compared with each other and with the native Dutch population. Several chapters provide details on the urban developments and social imbalances in the cities of the EU Urban Audit, on the complex relationships between cities and their peri-urban areas, and on commuting. The 2001 Census is compared with earlier Dutch population censuses from as far back as 1795. Furthermore, the Dutch Census is compared with the Census results of eight other European countries. Data from many different sources were combined to produce the Dutch Census Tables of 2001. Since the last census based on a complete enumeration was held, in 1971, the willingness of the population to participate has fallen sharply. Statistics Netherlands has found an alternative in the shape of the Virtual Census, using a combination of available registers and surveys. This puts the Netherlands in a unique position in the European Census Round. The Virtual Census is cheaper, comparable with earlier Dutch censuses, and more socially acceptable. In the last decade Statistics Netherlands has acquired more and more experience with using data from various administrative registers for statistical purposes. This has enabled Statistics Netherlands to develop the Social Statistical Database (SSD). The SSD contains coherent and detailed demographic and socio-economic statistical information on persons and households, for which the population register forms the backbone. In addition, sample surveys are used for information that is not available from registers, such as education level and occupation. The SSD is the main source for the 2001 Census. The Dutch Virtual Census of 2001 7 Compared with the 1991 Census Round, Eurostat and other international organisations required more detailed information for 2001. To meet the requirement of overall numerical consistency in the set of census tables, statistical methodologists at Statistics Netherlands developed a new estimation method: repeated weighting. It is based on the repeated application of a regression method, and it guarantees that the combination of survey and register information results in consistent estimates. The method is discussed in the last chapters of this book. Classifications, definitions and all census tables can be found on the website of Statistics Netherlands 2). The book is intended for researchers at universities, research institutes or statistical agencies, policy makers and the interested general public. I would like to thank the editors Rita Gircour, Marijke Hartgers and Eric Schulte Nordholt for their contribution. G. van der Veen, Director-General of Statistics Netherlands A Voorburg, July 2004. 1) Special thanks to Jessica Chamberlain from the Office for National Statistics in the UK for her contribution to chapter 12. 2) http://www.cbs.nl/en/publications/articles/General/census–2001/census–2001.htm 8 Statistics Netherlands 1. Introduction to the Dutch Virtual Census of 2001 Eric Schulte Nordholt Data from many different sources were combined to produce the Dutch Census Tables of 2001. Since the last census based on a complete enumeration was held in 1971, the willingness of the population to participate has fallen sharply. Statistics Netherlands found an alternative in the Virtual Census, using available registers and surveys. The Virtual Census is cheaper, comparable to earlier Dutch censuses, and more socially acceptable. The Netherlands takes up a unique position in the European Census Round. The table results are not only comparable with the earlier Dutch censuses but also with those of the other countries in the 2001 Census Round. The part on analysis in this book deals with the following topics: key figures, families, working people, young people, seniors, foreign people, and commuting. A historical comparison with earlier Dutch censuses is made and regional distributions are discussed with special attention to ten major cities in the Netherlands. Finally, the results of the 2001 Census in the Netherlands are compared with the results in other European countries. The part on methodology deals with the input, throughput and output phases placing special emphasis on how the new methodology of repeated weighting was applied in producing the set of census tables. 1.1 Introduction In 2003 data were combined to produce the Dutch 2001 census tables. In the Netherlands this was not done by interviewing inhabitants in a complete enumeration, but by using data that Statistics Netherlands already had available. This way, the Dutch tax payer got a much lower census bill. The costs for a traditional census would be about three hundred million euros, while the costs made now are ‘only’ about three million. The estimate includes the costs for all preparatory work such as developing a new methodology and accompanying software. The costs of the registers are not included, but the analyses of the results are. Registers are not kept up-to-date for censuses but for other purposes. Saving money on census costs is only possible in countries that have sufficient register information. As an example we can compare the costs of the Dutch virtual census with the costs of the traditional census that was held in Canada. In Canada the census costs amounted to approximately 450 million euros. Canada has about 31.6 million inhabitants, twice as many as the Netherlands. Statistics Canada justifies the huge census costs by pointing out the enormous implications of the The Dutch Virtual Census of 2001 9 census results for the distribution of money among regions. Moreover, a virtual census would be impossible in Canada because of the lack of sufficient register data. The 2001 Census relates to forty extensive tables. Twenty-eight are about the Netherlands as a whole, nine are at the COROP level (NUTS 3) and three at municipal level (NUTS 5). The forty tables fall into a number of groups. Eight tables concern housing, two tables concern commuting and the other thirty tables are demographic tables, relating to occupation, level of education and economic activity. Additionally, demographic, housing and labour figures are compiled at sub-city district level for ten large cities that participate in Urban Audit II (Statistics Netherlands, 2003). These ten large cities are Amsterdam, Rotterdam, The Hague, Utrecht, Eindhoven, Tilburg, Groningen, Enschede, Arnhem and Heerlen. Except the financial aspect, other important differences exist between a traditional census and the virtual census conducted in the Netherlands. In spite of the mandatory character of a traditional census, a certain part of the population will not participate (unit non-response) and the part that does participate will not answer some questions (item non-response). Correcting non-response by weighting and imputation techniques is well worth trying. A well-known problem with traditional censuses is that participation is limited and selective. Traditional correction methods fall short of the need to be able to publish reliable results. The last traditional census in the Netherlands (in 1971) met with much privacy objections against the collection of integral information about the population living in the Netherlands.