Merja Myllymäki, Statistics Finland, merja.myllymaki@stat.fi

Foreign trade and its significance for the national economy have long been measured by monitoring the development of trade in goods. Statistics on exports and imports of services have complemented the overall picture and the concept of economic ownership has clarified the principles of statistics. However, these alone are not enough to measure the importance of foreign trade, as traded products may consist of inputs produced in several countries. In reality, what is Finland's share in all this? To illustrate that, we needed regularly produced information on trade in value added, which draws attention to the domestic value added of exports.

Building on the previous efforts and collaboration of Nordic NSIs, and on the work done by OECD and WTO, Statistics Finland, together with the OECD, developed national statistics on trade in value added. By combining existing data in new ways, without any new data collection or additional response burden, we created statistics that give more accurate picture of the integration of different types of Finnish companies into global value chains. With the new statistics, it is possible to shed light on questions, such as, how the foreign trade actually benefits the national economy, and to what degree the labour force is dependent on global value chains.

Trade in Value Added statistics are published annually as part of the Statistics Finland’s Experimental Statistics. The paper describes the development work carried out with the OECD and the current production process in Statistics Finland and presents the selected indicators and company classifications. We will highlight key results, discuss limitations and statistical development points.

Keywords: trade in value added, global value chains, experimental statistics, new statistics – no response burden

Session: 1.1. Improving economic statistics

When: Tuesday, August 23 at 13:00 - 13:55

Download presentation

Fredrik W. Andersson, Statistics Sweden, fredrik.andersson@scb.se

To obtain a comprehensive picture of the population's status of employment, it is necessary to develop methods which identify individuals' who are self-employed and classify their status of employment. A problematic group are hybrids, i.e., those individuals who are at the same time both employees and self-employed. It is neither new nor controversial that self-employed underreport their business income. However, to produce statistics we can allow them to have only one status of employment: i.e., employed or self-employed. But how do you evaluate self-employed individuals and their business income in contrast to their income from an ordinary employment position?

The previous studies are based on the fact that Engel curves are offset in parallel between employees and entrepreneurs. This study uses the same theoretical approach, but instead of studying the data on food consumption, register data on interest expenses and holiday home ownership are used instead. All measures have their problems and additional study in this area is beneficial to conclude the level of tax evasion. This model's estimates provide very equal levels of previous Swedish studies’ estimated underreporting for the Swedish population of self-employed.

The result indicates that self-employed people in unincorporated business underreport about 50 percent of their business income, while the underreporting for self-employed in incorporated business is between 15 and 20 percent. Hence, underreporting is an aspect which is needed to be considered for an unbiased estimation of status of employment.

Keywords: Self-employed; status of employment; administrative data; novel approach.

Session: 1.1. Improving economic statistics

When: Tuesday, Tuesday, August 23 at 13:00 - 13:55

Download presentation

Elise Tengs, Statistics Norway, elise.tengs@ssb.no

Einar Døvik Stavnes, Statistics Norway, einar.stavnes@ssb.no

Leiv Solheim, Statistics Norway, leiv.solheim@ssb.no

Response burden management is important for all National Statistical Agencies (NSI) in their work to reduce the cost inflicted on respondents. The Fifth Principle of the United Nation’s Fundamental Principles of Official Statistics (2014) demands that statistical agencies consider “the burden on respondents” when collecting data. An important way in which many NSIs ensure minimal response burden in business surveys is through a probability sampling coordination system. This system draws representative samples of businesses for each survey, while at the same time minimizing the overall number of surveys each business has to participate in over a given period. However, the number of surveys is not the only way to measure response burden. The time spent on answering the survey, and the overall user experience, are also important measures of response burden.

This paper explores ways in which this information – time spent, and user experience – can be accounted for when designing a system for sampling coordination of business surveys.

Statistics Norway (SSB) uses the coordination system called NORSAMU (Norwegian System for Coordinated Business Surveys). Also, SSB collects information on time spent and user experience through the business survey portal called Altinn. However, this information from Altinn is yet to be accounted for within NORSAMU. This paper analyses the total response burden for each business, when combining information both from NORSAMU and from Altinn. Importantly, this paper answers the question: How can data on time spent and user experience be implemented when designing a system for sampling coordination of business surveys, in order to improve response burden management?

This will provide new insights relevant for other NSIs in their response burden management within sampling coordination of business surveys. This analysis also provides an opportunity for exchanging ideas between countries with similar systems.

Keywords: business surveys, probability sampling coordination, response burden management

Session: 1.1. Improving economic statistics

When: Tuesday, Tuesday, August 23 at 13:00 - 13:55

Download presentation

Martin Axelson, Statistics Sweden, martin.axelson@scb.se

Susanne Gullberg Brännström, Statistics Sweden, susanne.gullbergbrannstrom@scb.se

As part of the strategy adopted by Statistics Sweden in 2019, four endeavours are identified. One of these focuses on modernising and future-proofing the labour market statistics. As an important part of this endeavour, a development initiative labelled Subject Area Design Labour Market was conducted in 2019-2021. The work aimed to analyse and identify current use and needs of labour market statistics, and using this as a starting point, to provide a broad analysis of the status of these statistics. The task also included proposing new statistics that describe the Swedish labour market. The work was conducted in dialogue with central users and was carried out in cooperation with Swedish Public Employment Service, the Swedish Work Environment Authority, and the National Mediation Office.

In May 2021, Statistics Sweden presented a three-stage plan for its future labour market statistics. The plan forms part of the agency’s stated ambition to meet user needs and at the same time reduce its dependence on direct collection of data. The statistics are to meet user needs and must be of good quality, while taking into account aspects such as resource use and respondent burden. An important starting point in preparing the plan is to use, as far as possible, already collected data for the production of the statistics. Special importance is given to administrative data from other government agencies, the goal being that future statistics are to be register-based and sample-assisted, rather than vice versa.

The first part of the paper discusses the initiative Subject Area Design Labour Market and presents the main components of the plan. The second part of the paper provides information on Population by Labor Market Status, a new, register-based product which marks the first step towards implementation of the plan. The paper is concluded with reflections on the work carried out so far and the way forward.

Keywords: user dialogue, design

Session: 1.2. New data sources

When: Tuesday, August 23 at 14:00 - 14:55

Download presentation

Pernille Stender, Statistics Denmark, psd@dst.dk

Susanne Gullberg Brännström, Statistics Sweden, susanne.gullbergbrannstrom@scb.se

The Öresund region is Northern Europe’s largest labour market region with approx. 4 million inhabitants. After the opening of the Öresund Bridge in 2000, cross-border commuting in the region increased considerably. Therefore, there was a great demand for statistics that could shed light on commuting flows between the Danish and Swedish parts of the region. As a consequence, Statistics Denmark and Statistics Sweden developed statistics on cross-border commuting which was released for the first time in 2000. The statistics (the old cross-border statistics) were based on the exchange of data between Statistics Denmark and Statistics Sweden. The statistics published in 2000 had information about cross-border commuting in November 1998. Later, figures for 1997 were also published. Thereafter, the statistics were produced yearly until the release of figures for 2015 in 2017. Since the publication of figures for 2015, it has not been possible to update the statistics. The statistics can no longer be produced due to changes in legislation, which imply that it is no longer possible to exchange microdata at personal level between the two statistical agencies. This has been of great inconvenience to decision makers, analysts etc.

In the second half of 2021, Statistics Denmark and Statistics Sweden carried out a feasibility study. The study was funded by Region Skaane. The aim of the study was to investigate whether it would be possible to re-establish cross-border commuting statistics for the Öresund region. The premise of the study was that it should be possible to produce the statistics without data exchange between the statistical offices.

Keywords: nordic-cooperation, cross-border statistics

Session: 1.2. New data sources

When: Tuesday, August 23 at 14:00 - 14:55

Download presentation

Sanna Roos, Statistics and Research Åland, sanna.roos@asub.ax

It all begins with an idea. Maybe you want to launch a business. Maybe you want to turn a hobby into something more. Or maybe you have a creative project to share with the world. Whatever it is, the way you tell your story online can make all the difference.

The main theme of the paper refers to the challenge of generating reliable data and statistics on alcohol consumption at the regional level. The need for regional estimations is emerging as alcohol policies aim at targeting consumption in different subgroups of the local population, such as among youth and elderly. Since the 1970s, estimation of total alcohol consumption in population studies have largely been rooted in the theory of the total consumption model (TCM). At the same time, studies on alcohol consumption at the regional level have dominantly been dependent on survey methods, which are known to be challenged, for instance, by the growing non-response rates. Furthermore, local and contextual factors need to be evaluated when designing and building data collection aiming at serving the needs of regional alcohol policy formation. In Åland, the easy access to duty-free alcohol contributes to the total alcohol consumption at the local level, at the same time as the lack of statistics on alcohol purchases made by the tourists severely hampers the estimation of alcohol consumption in the local population. Thus, innovative ways are needed in collecting and quantifying data and statistics on alcohol consumption in local populations. This is highly important, in particularly, in those Nordic regions which might be facing specific challenges due to the unique contexts. Recommendations for supplementary methodologies and actions are presented.

Keywords: total consumption model (TCM), alcohol policy, surveys, Åland

Session: 1.2. New data sources

When: Tuesday, August 23 at 14:00 - 14:55

Download presentation

Maria Pedersen, Statistics Denmark, mrp@dst.dk

Paul Lubson, Statistics Denmark, pal@dst.dk

This paper examines the Danish statistics on library services on an individual level. The paper outlies the cooperation established between Statistics Denmark and the public library sector in Denmark. The end result is automated daily reports provided from the public libraries to Statistics Denmark containing information on an individual level linking users with their library borrowing. The setup with regards to handling data is discussed as well as how privacy of the users is ensured. Data covers both the traditional public libraries and the digital ones.

Next, information on the users is combined with demographical statistics in order to profile the user. Age, sex, education, income and social status are among the demographics linked to the library user and the end result is a detailed profiling of the behavior of different demographic segments. This data is produced both on a national level and on a municipal level. Furthermore, the address of each library user is used to profile the user base of each public library. We profile both their local user demographics, but also through interference the demographics of the locals that are not library users.

Finally, examples of how the daily data mapped use of libraries during the pandemic are shown.

Keywords: Library statistics, Demograpic profiling. Automated big data.

Session 1.3. Data collection challenges I

When: Tuesday, August 23 at 15:30 - 16:25

Download presentation

Nina Berg, SSB, Nina.Berg@ssb.no

Gezim Seferi, SSB, Gezim.Seferi@ssb.no

Contributors: Aina Holmøy, Kristin Egge-Hoveid, and Kari-Anne Lund, Statistics Norway

The Norwegian Household Budget Survey 2022 is the first survey to utilize mobile app for data collection at SSB. This survey will be used as a case: 1) to describe how we developed the survey instrument, eg the mobile app, as well as our survey communication for data collection, and 2) to evaluate the results in respect to response rates and sample composition.

A main goal in our evaluation is to look at the effects of digitalisation through app usages and respondents’ participations.

• Will the format and the design of the instrument include or exclude some respondent groups?

• Will this digitalisation of data collection decrease, increase, or maintain adequate response rates?

• Will this digitalisation of data collection decrease, increase, or maintain adequate composition of different respondent groups?

We have both qualitative and quantitative data to evaluate this case. And we will look at follow up interviews with respondents or participants of the survey. We will have multiple paradata or process data and survey data from first quarter/half year in our analysis. F example:

• Respondent data on activity of registration - when and frequency of response, like day and time of day, and number of registrations.

• Back end data on survey communication, such as when e-mails and follow up Short Message Service (SMS) are sent.

• Response rates.

• Qualitative and quantitative data on reasons for none response. Particular in regards to digital skills.

We will summarize what we have learned and discuss possible best practice when using mobile apps for data collection.

Keywords: Digital skills, mobile app, survey communication, response rate, response burden

Session 1.3. Data collection challenges I

When: Tuesday, August 23 at 15:30 - 16:25

Download presentation

Thomas Laitila, Statistics Sweden, thomas.laitila@scb.se

Low response rates are generally obtained in household expenditure surveys collecting data via diary keeping. It is reasonable to question the quality of resulting estimates of population characteristics. However, low response rates do not necessarily mean invalid estimates and the opposite may be equally true. This first may be the case particularly for surveys involving diaries, where the respondent records events during a time interval ahead. This implies the study is prospective meaning the choice of responding is made prior to recording events. Thus, what is to be observed cannot be assumed intervening the choice to respond or not. Dropouts during the data collection period may on the other hand be directly affected by realized values on study variables.

The prospective character of diary surveys is detailed upon showing the possibility to derive valid estimates using auxiliary variables. A main issue addressed is what auxiliary variables to use and how to use them in estimation. It is suggested economic and behavioral theory to be used for the modelling of response probabilities. An illustration with an example from the Swedish household survey is presented, and includes a model derived from economic theory suggested for modelling of response probabilities. Finally, the importance of having the response set covering the range of population values of study variables is shown. The implications of results for the design of diary-based surveys are discussed.

Keywords: diary survey, prospective studies, response probability, auxiliary variables

Session 1.3. Data collection challenges I

When: Tuesday, August 23 at 15:30 - 16:25

Download presentation

Mona Eide Onstad, Statistics Norway, Mona.Onstad@ssb.no

Manju Chaudhary, Statistics Norway, Manju.Chaudhary@ssb.no

Gisle Berge, Statistics Norway, Gisle.Berge@ssb.no

With an increasing focus on reuse and circular economy in the EU and UN, there is a demand for such statistics. Through the Directive of the European Parliament and of the Council (EU) 2018/851, Norway is required to report on the reuse of textiles. To meet this growing demand from both a society concerned with climate change, and from the EU and the UN, we began investigating how we can use online sources to find data on textile reuse in Norway.

Our perception is that a considerable portion of the market of reuse is online, through web pages such as Finn.no, Tise, Fretex and Facebook Marketplace. Any analysis on reuse would be incomplete without the inclusion of online platforms offering exchange and reuse of already used items. With this being the case, a lot of data should be readily available for us to employ in a potential statistic on reuse. The objective of this paper is twofold. First, we investigate the accessibility of the data – is it as readily available as it appears? We find that quick access solutions such as web scraping have too many limitations, and that establishing contact with the online platforms is preferable when possible. Data can then be received through files or using application programming interface (API). The second objective of the paper is to assess the quality of the data that we can gather. Specifically, whether it is usable and complies with our needs for official statistics.

As more data is becoming available through new, online sources, it is important to research whether it can lessen the work load that goes into surveys and other methods for establishing new statistics and can be a good quality data source. In this paper we share the experiences that we have gained, dipping our toes in this new area.

Keywords: reuse, online sources, digital platforms, circular economy

Session 1.4. Data collection challenges II

When: Wednesday, August 24 at 10:30 - 11:25

Download presentation

Jens Malmros, Statistics Sweden, jens.malmros@scb.se

Jannis Kalpouzos, Statistics Sweden, jannis.kalpouzos@scb.se

The Living Conditions Survey of Children (Children’s LCS) is a survey of children 12-18 years old given by Statistics Sweden since 2001. Survey topics include economy and material resources, health, and social relations. The Children’s LCS was previously a supplementary survey to the Living Conditions Survey (LCS). Because of declining response rates and changes in the LCS design, the Children’s LCS will turn into a stand-alone survey.

Because most surveys of children are supplementary surveys or classroom surveys, knowledge of surveys of children with direct element sampling is limited. For example, there is substantial uncertainty on data collection method, response rates, and representativity of responding sample persons. To address this, Statistics Sweden carried out an experiment on the choice of data collection method for the stand-alone Children’s LCS. In the experiment, mixed mode using telephone and web interviews were compared with web interviews.

The response sets from the two data collection methods were evaluated with respect to desired quality and granularity. For example, response rates and the expected number of respondents in domains were compared. In addition, the bias and variance of estimates of survey variables and estimates of register variables were evaluated. The overall response rate was 56 % in mixed mode and 38 % in web.

Both methods showed similar results with respect to the evaluated measures when adjusted for the difference in overall response rate; hence, the conclusion is that both methods result in adequate quality for the Children’s LCS. Because of lower data collection costs, it is possible to choose a much larger sample size in web for the same overall cost, which results in a larger response set compared to mixed mode. It is therefore recommended to use web as data collection method for the Children’s LCS. Statistics Sweden recommends conducting proxy interviews by telephone.

Keywords: Survey of children, data collection, experiment, mixed mode, web

Session 1.4. Data collection challenges II

When: Wednesday, August 24 at 10:30 - 11:25

Download presentation

Ingrid Hatlebakk Hove, Statistics Norway, ingrid.hatlebakk.hove@ssb.no

Frode Berglund, Statistics Norway, frode.berglund@ssb.no

Borgny Vold, Statistics Norway, borgny.vold@ssb.no

Changing environments raises new needs for statistics. Because of the pandemic, there is a huge interest in society for information concerning intensive care capacity in hospitals. In 2020, Eurostat for the first time asked the European countries to report the number of intensive care beds.

In Norway, there was no existing official statistics on intensive care beds. Statistics Norway had to respond to the new requirements and consider the possibilities to create official statistics on intensive care capacity.

Co-workers from the health statistics division teamed up with a co-worker from the division for methods. Together we raised several questions: Could we include questions on intensive care in the existing form we already send to the hospitals annually? If so, is it possible for the respondents to report the numbers according to Eurostat’s definition?

The easiest option would have been to simply translate Eurostat’s requirements into questions, send directly to the hospitals and create statistics based on the data we collected. However, we didn’t know whether the required information was easily available for the respondents. Instead, we arranged online meetings with the hospitals to discuss the challenges and possibilities of reporting numbers of intensive care beds.

Collecting the data according to Eurostat’s definition turned out to be challenging. In this article we share our experiences and highlight the importance of discussing with respondents before starting a new data collection. Only by talking to the respondents we can find out which data we might collect, how to ensure that we get the data we think we get, and how to minimize the response burden.

Keywords: complexities in data collection, respondents, intensive care

Session 1.4. Data collection challenges II

When: Wednesday, August 24 at 10:30 - 11:25

Download presentation