Chair: Orietta Luzi
Room: S3B Sukiennice
Time: 13:15 - 14:00
Date: 27 June
Speed Talk Session 3 - papers & presentations
|Title: <<< Integration of New Administrative Data Sources into Turkish Statistical System >>>
Turkstat aims to collect and process data more quickly, more accurately and more effectively than before. On the other hand; unlike the past, it seems like data collection works have been changing rapidly in recent years. Emerging technologies, respondent burden concerns, increasing costs, globalization, having more difficulties in collecting data than the old times and exponentially increasing amount of data day by day and thus increasing data demands are among the reasons for these changes. Overcoming the data collection challenges, Turkstat has given priority to "increasing the weight of administrative registers in statistical production" in its future plans and programs and has accelerated the studies on this subject. Turkstat started to pay more attention to take advantage of administrative sources. In this context, TurkStat started regular monthly data transfers from Revenue Administration as of June 2017 and from Social Security Institution as of October 2017. Starting from 2018, as well as structural business statistics characteristics, short term business statistics indicators such as turnover, production and labor input are planned to be produced by using these administrative records. Furthermore, as data collection shifts to secondary sources due to integration of new data sources, Turkstat started to have a new employee profile having combined capabilities of statisticians and IT specialists. Making use of new administrative data collection activities in Turkstat also started to cause developing new skills, creating a mindset and culture open to change and developing strategic partnerships with administrative registers holding public institutions. This paper shows Turkstat’s experiences on integration of new administrative data sources into Turkish Statistical System, the resulting necessity of redesigning its business statistics and analytical skills and competencies developed in handling large and complex administrative datasets.
|Title: <<< Managing changes in key administrative data sources within Finnish SBR >>>
The latest Eurostat Code of Practice (CoP) peer review was organized by Eurostat in 2013–2015. The Finnish country report (2015) identified, among other findings, a strong reliance on administrative data sources in statistical production. Recommendation #6 of the report called for risk assessment of possible changes in these administrative sources. The objective of the so-called Valmis-project (2013–2019) is to develop operations and reform taxation software. Numerous taxation software will be replaced by one software package (GenTax). These changes will affect the administrative data that the National Board of Taxes provides to Statistics Finland. There will be changes to both the content and the format of the data. At the end of 2015 the Administrative Data Collection team (AVa), part of the Data Collection Unit at Statistics Finland, prepared a report for The (National) Statistical Law Working Group outlining risks and risk management with the use of administrative data sources in statistical production. Key points of this paper include:
- An outline of the Tax administration's Valmis project.
- Analysis of the effects the Valmis project on The Statistical Business Registry (YTY).
- Assessment of the centralized input of data and change management procedures of YTY in adapting to these changes.
- Assessment of the benefits gained from risks management plans for identified administrative data sources.
- Analysis of methods to benefit quality control from materialized risk scenarios.
- Lessons learned from a materialized risk situation: a uniform abstract layer between data input and the Statistical Business Registry.
|Daisie Hutchinson |
|Title: <<< Harmonisation across the UK: Comparability of survey and administrative data in conjunction with European influences >>>
The Harmonisation Team work across the UK Government Statistical Service (GSS) by harmonising definitions, questions and outputs to ensure that the quality of official statistics meets user requirements and supports comparability both within the UK and internationally. The GSS harmonisation vision is that all definitions, questions and outputs for the census and surveys and all data from administrative records will be harmonised, so that users can compare data from different sources with confidence and can merge and match data more easily. The harmonisation of administrative data is a new and challenging area of work and forms a large part of this harmonisation work programme. Understanding the conceptual differences between survey data and administrative data is key to this. The harmonisation team plan to develop a set of harmonised definitions, questions and outputs for administrative data for use across the GSS. This will be in addition to the harmonised definitions, questions and outputs already in place for surveys. A further new area of work is the harmonisation of business statistics to comply with EUROSTATs Framework Regulation Integrating Business Statistics (FRIBS). The regulation requires the GSS to move to a harmonised set of variables by 2019. Alongside this, and in collaboration, ONS are harmonising business survey questions where possible on the Electronic Data Collection programme over the next few years. The paper sets out what the GSS has achieved to date with harmonisation, including the development of a harmonised question library, harmonised definitions, questions and outputs for surveys, administrative data and also what remains to be done. It also outlines the benefits of harmonising and details the issues and challenges faced when attempting to harmonise.
|Title: <<< Record linkage in agricultural statistics >>>
The use of data from administrative registers have been used extensively in Sweden since Sweden became a part of the European Union in 1995. Integrating administrative registers with censuses and sample-surveys has been seen as a cost-effective way of producing statistics with sufficient quality. The integration phase where data from several sources is integrated into a new statistical register is seen as essential for achieving sufficient quality. To successfully link records from a specific unit in an administrative registers with a corresponding unit in a statistical register is therefore essential for the quality of the final statistics. In some cases, the linkage is perfect but in many cases, the unit in the administrative record do not uniquely relate to the unit in the statistical register. Choices and rules taking into account the information at hand must then be used to perform the record linkage. In this article, the outcome of using different rules for linking data from administrative registers into the statistical farm register is discussed.
|Title: <<< Data integration to quantify occupational injury risk wage premium >>>
Studies of occupational injury risk influence on wages and wage differentiation are rare due to methodological and data problems. In this study we show how to take advantage of the surveys conducted within Polish Statistical Surveys Programme while analysing occupational risk. We combine information from two public data sources: the Structure of Earnings Survey (SES) and the Accidents at Work Data regained from the Statistical Cards on accident at work to estimate the influence of occupational injury risk premium on wages and the gender wage gap. While the SES is a well-known database among economists analysing labour market, the latter is a less popular one: it contains annual data about all of the accidents and their characteristics. Merging of the databases enables us to operate on the individual level instead of aggregated data. We match every observation from the SES data with estimated accident rates calculated separately for every economic section/occupation/gender cell. We show that it is feasible to work at three digit level of the International Standard Classification of Occupation and still obtain precise estimates of occupational injury risk for both genders. Contrary to the existing research, we take directly into account the differences in work related risk among workers. As an illustration, we use the combined data in occupational injury risk wage premium estimation, and answer the question to what extent the occupational injury risks explain the existing gender wage gaps.
|Title: <<< Data integration - idea for reduction of complexity >>>
MONSTAT conducts hundreds of surveys every year and prepares dissemination covering virtually every aspect of statistics in the Montenegro. Our mission is to provide timely, accurate, and useful statistics in service to the government of the Montenegro. Despite complexity of surveys and the large volume of data processed daily, the most important goal is to provide precise, timely and reliable data. The most usual way of processing data of a statistical survey is a, so-called, stovepipe principle, where the complete process is performed – from data entry to publishing the results separately for each survey. Therefore, NSIs have a situation where the processing of statistical data is done on various platforms, implemented on a variety of software tools, with data stored in various ways, separately for every survey, and even worse, for every statistical phase. Many of NSIs overspend on technology in the quest for getting value for the institution. They need to manage large IT structures, which are very expensive and slow, in order to approve capital investment and operational costs. One of the key distinguishing factors of statistical offices is data integration. Other organizations and enterprises may have better insight in individual sources but NSIs know best how to combine data and turn them into reliable figures describing the whole society. We gradually put all data on one platform. In that way we got all the functionalities of a warehouse. We made a uniform approach to all surveys, put all information about every piece of data in one place and made a single tool to handle all that information and data. That is IST Integrated system of data processing. IST was designed, developed and implemented in four NSIs and present example of international collaboration in data integration area.