New Possibilities for Socio-Economic Research through Longitudinal Data from the Research Data Centre of the German Federal Pension Insurance (FDZ-RV)

Himmelreicher, Ralf K.; Stegmann, Michael Article New Possibilities for Socio-Economic Research through Longitudinal Data from the Research Data Centre of the German Federal Pension Insurance (FDZ-RV) Schmollers Jahrbuch – Journal of Applied Social Science Studies. Zeitschrift für Wirtschaftsund Sozialwissenschaften Provided in Cooperation with: Duncker & Humblot, Berlin Suggested Citation: Himmelreicher, Ralf K.; Stegmann, Michael (2008) : New Possibilities for SocioEconomic Research through Longitudinal Data from the Research Data Centre of the German Federal Pension Insurance (FDZ-RV), Schmollers Jahrbuch – Journal of Applied Social Science Studies. Zeitschrift für Wirtschaftsund Sozialwissenschaften, ISSN 1865-5742, Duncker & Humblot, Berlin, Vol. 128, Iss. 4, pp. 647-660, https://doi.org/10.3790/schm.128.4.647 This Version is available at: https://hdl.handle.net/10419/292246 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ European Data Watch This section offers descriptions as well as discussions of data sources that are of interest to social scientists engaged in empirical research or teaching courses that include empirical investigations performed by students. The purpose is to describe the information in the data source, to give examples of questions tackled with the data and to tell how to access the data for research and teaching. We focus on data from German speaking countries that allow international comparative research. While most of the data are at the micro level (individuals, households, or firms), more aggregate data and meta data (for regions, industries, or nations) are included as well. Suggestions for data sources to be described in future columns (or comments on past columns) should be send to: Joachim Wagner, Leuphana University of Lueneburg, Institute of Economics, Campus 4.210, 21332 Lueneburg, Germany, or e-mailed to hw[email protected]i. Past “European Data Watch” articles can be downloaded free of charge from the homepage of the German Council for Social and Economic Data (RatSWD) at: http://www. ratswd.de. New Possibilities for Socio-Economic Research through Longitudinal Data from the Research Data Centre of the German Federal Pension Insurance (FDZ-RV) By Ralf K. Himmelreicher and Michael Stegmann 1. Introduction For many socio-economic analyses micro-data on the life-course of the individual or at least on the employment-biography is needed. The reason for this is that micro structures and developments in socio-economics can only be analysed in detail on the basis of individual data. The dynamics behind changes in micro-economic activities can only be depicted with longitudinal data. Longitudinal data on individuals is also needed in order to analyze socioeconomic aspects of the developments in the life-course (Schmähl / Fachinger, 1994, 180 ff.). An example is the evaluation of the question whether higher Schmollers Jahrbuch 128 (2008), 647– 660 Duncker & Humblot, Berlin Schmollers Jahrbuch 128 (2008) 4 OPEN ACCESS | Licensed under | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.128.4.647 | Generated on 2023-01-16 13:35:36 648 Ralf K. Himmelreicher and Michael Stegmann education results in higher life-course earnings. Finally, there is an aspect that longitudinal data allows the calculation of models which are superior to models with cross-sectional data, for example with regard to their explanatory power (Engel / Reinecke, 1994, 5 ff.). The Research Data Centre of the German Federal Pension Insurance (FDZRV) supplies cross-sectional and now also longitudinal micro-data. This data is process-produced. In Germany statutory pension insurance is mandatory for all employed persons in the private and public sector. In additional to this, contributions are paid out of unemployment insurance in the case of the unemployed and out of health insurance in the event of long-term illness. So the majority of the German population comes into contact with the German Federal Pension Insurance and the pension data covers more than 90 % of the entire population (Rehfeld / Mika, 2006, 121 ff.) The German Federal Pension Insurance is the biggest income source after retirement. Besides old-age pensions another important function is rehabilitation in the event of inability to work. So the FDZ-RV also provides information on health status data. The second chapter of this paper gives an overview of the data and especially of the longitudinal data of the German Federal Pension Insurance. The third shows how the FDZ-RV creates Scientific Use Files (SUFs) from genuine process-produced data and gives some examples of research topics. Finally we make some remarks on the FDZ-RV, in particular on the data-supply in the present and the future. 2. Data of the German Statutory Pension Insurance Figure 1 gives an overview of cross-sectional and longitudinal micro-data of the German Federal Pension Insurance in the fields of retirement, insured persons and rehabilitation with the corresponding names of the micro-data. This grey-marked data is generated from the FDZ-RV as anonymized SUF 1 , which you can get free of charge as a scientist working in a research unit, the only requirement being a signed contract with FDZ-RV. The statistics of the German statutory pension insurance can be divided into datasets that focus on biographic information in combination with retirement and insurance and in a special dataset for rehabilitation. The data-sets listed with a reference period of one day means that this day represents the monitoring day in a specific year. Some statistics have both reference periods, daily and yearly. The longitudinal dataset for rehabilitation is mentioned here to give a complete overview of the existing data sources. The preparation of this data in the Schmollers Jahrbuch 128 (2008) 4 1For general rules of de facto anonymization in the FDZ-RV see Stegmann et al. (2005). OPEN ACCESS | Licensed under | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.128.4.647 | Generated on 2023-01-16 13:35:36 New Possibilities for Socio-Economic Research 649 form of Scientific Use Files has not yet been implemented. For this reason we will only look at this topic briefly. This article focusses on the longitudinal datasets on new pensioners and insured persons. Process-produced longitudinal data of social security is especially useful because the employment career can be analysed. Therefore typical problems with longitudinal surveys in social research can be avoided. Panel mortality and recall errors concerning retrospective questions cannot occur. Source: Following Himmelreicher / Radl (2006). Figure 1: Micro-data-sets of the German Federal Pension Insurance 2.1 Longitudinal Data-sets on Insurance and Retirement Information on insured persons who are registered in the state pension insurance records is the basis for the statutory pension calculation. Pension-relevant creditable periods are recorded. From the age of 17 (earlier, if a contribution has been paid) until retirement, all contribution periods and relevant non-contributory periods are registered, giving information on the living conditions in a special period of time, and showing periods of insurance liability or periods of child-raising. Schmollers Jahrbuch 128 (2008) 4 OPEN ACCESS | Licensed under | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.128.4.647 | Generated on 2023-01-16 13:35:36 650 Ralf K. Himmelreicher and Michael Stegmann Source: Following Stegmann (2007: 22, Figure 1). Figure 2: Longitudinal micro-data-sets of the German Federal Pension Insurance By means of the insurance account sample, the so-called “Versicherungskontenstichprobe (VSKT)” and the so-called “Vollendete Versichertenleben (VVL)”, two special data-sets could be differented concerning the fields of insurance and retirement. These two data-sets with reference to longitudinal information have predominantly the same structure but a different category of persons. The VSKT provides information on round about 240,000 insured persons of German Statutory Pension Insurance aged between 15 and 67 years and their pension entitlement. Information is given on all relevant registered creditable periods. The results are the basis for the planning department, also for internal planning in insurance. Information is also used in order to inform the press and the public. Statistics are intended to make information available in order to support the preassessment of to the financial development of pension insurance. Schmollers Jahrbuch 128 (2008) 4 OPEN ACCESS | Licensed under | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.128.4.647 | Generated on 2023-01-16 13:35:36 New Possibilities for Socio-Economic Research 651 Concerning the insurance accounts the reporting is done in a kind of a random sample and is carried through in the next years as a panel. A panel structure has been chosen in order to draw on cleared insurance accounts. On the other hand it is important to minimize the expense because of the effort needed to insurance accounts. The category of insured persons in the statistics covers all persons who are registered in an insurance account of German pension insurance which –is not closed at the point of time of the evaluation –contains contribution periods till the census deadline 31.12. of the reference year, means: the insurance account is not empty or a bonus is registered as a result of settlement of pension entitlements –is without registration till the census deadline (31.12. of the reference year) applying to the notification of the death of the insured person –concerns a person of a minimum age of 15 (31.12. is census deadline of the reference year) and a maximum age of 67. The insured persons represent the entire population of the survey. In order to present the panel as a kind of random choice, all insured persons in the primary survey in the year 1983 and newly insured persons in the following years have to be registered proportionately. This occurs by restricting the sample concerning the newly added insurance numbers. In order not to distort the sample it is important to use the same survey probability in the case of post-operative sampling with reference to the occupied groups and the original primary panel survey. Post-operative cases are marked by being filed into the database in the year of the census deadline. The post-operative sampling occurs at the beginning of the year. The special survey “Vollendete Versichertenleben (VVL)” is to enable empirical evaluation of changes in pension law with respect to their impact. A sample is taken of the cases of pensions awarded in which the accounts are completely edited. For this sample a scale of 200,000 cases (round about 1 /5 of the total “inflow-sample” (Rasner et al., 2007,15) population) is strived for. In particular new recipients of insured persons‘ pensions and non-contract pensions are part of this survey. The VVL gives information on cases of pensions awarded within the reference year. Contract pensions and pensions awarded in which a supranational pension entitlement plays a role are excluded. The longitudinal data of the insurance accounts which are shown in the “VSKT” and the “VVL” are similar in their structure as to the preparation of data. In the case of “VVL” concerning the pensions awarded, the statistic data set also exist. The data-sets contain statistical characteristics and are established for all cases of pensions awarded. Schmollers Jahrbuch 128 (2008) 4 OPEN ACCESS | Licensed under | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.128.4.647 | Generated on 2023-01-16 13:35:36 652 Ralf K. Himmelreicher and Michael Stegmann 2.2 Longitudinal Data-Sets on Rehabilitation German statutory pension insurance is responsible for rehabilitative measures for insured persons and their relatives. In this connection the statistical reporting system services a special database that contains information for recipients of rehabilitative measures. This can be differentiated in medical and occupational rehabilitation. The basic aim of all rehabilitative measures is to restore the ability to work and avoid a persistent inability to work. In the field of rehabilitation, there is a special micro-data-set called RehabStatistik-Datenbasis (RSD). The RSD contains information on claims for rehabilitation (Anträge auf Leistungen zur Teilhabe), their execution and claims for retirement pensions. This data-base combines insurance and retirement information of either eight or eleven years for cross-sectional and longitudinal analysis. The RSD population is defined as such persons from the age of 52 to 66 who claim rehabilitation or retirement pensions in a reference year (see figure 2). 3. Longitudinal Data of the Research Data Centre (FDZ-RV) 3.1 Generating a Scientific Use File based on Longitudinal Register Data The Scientific Use File Completed Insurance Biographies 2004 (Vollendete Versichertenleben 2004, short SUF VVL 2004) is the first longitudinal dataset provided by the Research Data Centre of the German Federal Pension Insurance. 2 The SUF is based on administrative pension records of individuals, who are entitled to receive public pension benefits. The creation of the Scientific Use File VVL 2004 is an important step to enhance the usability of register data for researchers interested in issues of disability and retirement. The longitudinal data format enables researchers to analyze research questions related to the life-course, earnings and old-age income security. The SUF VVL 2004 is a systematic random sample of individuals who received state pension benefits for the first time in 2004. 3 In the first step, a 20 % sample was drawn from the pool of first-time retirees in 2004. Only persons receiving old-age pension benefits (Altersrente) and disability benefits (Erwerbsminderungsrente) can be part of the sample population. In the second step, a subsample of 25 % was drawn from selected age groups, namely persons born between 1939 and 1975. The final data product, the SUF VVL Schmollers Jahrbuch 128 (2008) 4 2For the remainder of the article, the authors refer to the short form of the dataset Completed Insurance Biographies 2004, namely SUF VVL 2004. 3The SUF VVL 2004 is a so-called inflow sample. Only inflows into retirement, whether old-age or disability benefits, can be part of the sample (Rasner et al., 2007). OPEN ACCESS | Licensed under | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.128.4.647 | Generated on 2023-01-16 13:35:36 New Possibilities for Socio-Economic Research 653 2004, is a 5% sample of first-time pensioners with a sample size of 39,331 cases. In the analysis of the data, it needs to be taken into account that the data is selective for numerous reasons. The selectivity is partly due to the sampling procedure applied and partly related to the legal rules of the state pension system. Firstly, only persons eligible for old-age or disability pension benefits can be part of the sample. Certain subgroups of the population are therefore systematically excluded. Among those are civil servants and the self-employed when they have never accumulated any pension entitlements in the social security system. Civil servants and the self-employed are also excluded if they do not fulfil the minimum qualifying period. 4 Secondly, certain benefit types are excluded from the sample, such as survivor’s or educational pension benefits. Thirdly, so-called Vertragsrentner agree excluded from the sample. Vertragsrentner are eligible for state pension benefits in a foreign country with which Germany has a social security agreement. As a consequence of these selection criteria, the SUF VVL 2004 sample population is representative neither of the population as a whole, nor of the group of the elderly. Hence, it is impossible to draw inference from the sample population for the entire population. For example, the distribution of old-age income in the SUF VVL 2004 does not give any indication of the prevalence of old-age poverty in the total population. Conclusions and generalizations can only be drawn for the respective sample population. The SUF VVL 2004 consists of two main components. The first part contains cross sectional variables (e.g., year of first-time pension receipt, gender, nationality, etc.), as well as aggregated data related to the calculation of the individual’s state pension benefit. The second component of the data is subdivided into several longitudinal files. Ideally, the longitudinal information is available for a maximum of 624 months, starting in January in the year the person turned 14 up to December in the year the person turned 65. A month is coded as missing if no pension-relevant information applies. Other longitudinal files contain information about the monthly earning points accumulated by each individual (file mEGPT) 5 , longitudinal information on whether a person worked in marginal employment (file NJOB) or worked as a caregiver (file PFLEGE), etc. 6 Longitudinal information is also provided on the social employment situation of an individual (for more information see below). Schmollers Jahrbuch 128 (2008) 4 4The minimum qualifying period amounts to five years of employment subject to social insurance contributions or 10 to 15 years if creditable periods or credited substitute periods are also considered. 5The earning point is one component of the pension benefit formula. The earning points reflect the earnings position of the individual relative to the earnings position of all persons paying social insurance contributions. 6For more detailed information on the available longitudinal data see Stegmann, 2006, 549 f. OPEN ACCESS | Licensed under | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.128.4.647 | Generated on 2023-01-16 13:35:36 654 Ralf K. Himmelreicher and Michael Stegmann Any analysis based on data provided by the Research Data Centre requires thorough knowledge of the German Social Code. All the available information is related to pension-relevant activities. For example, information on the birth of a child is only available in the data if the insured person is eligible for childcare credits. In turn, the birth of a child is not pension-relevant if a person works as a civil servant. 7 All data provided by the Research Data Centre needs to be analyzed in the light of the rules and regulations of the German Social Code. However, these rules are not constant over time. Hence, changes in the law are reflected in the data. For example, for all children born before January 1, 1992 childcare credits amount to one year. For all children born thereafter, three years are credited to the pension account of one parent. The interpretation of results based on the Scientific Use File is not straightforward, but results need to be interpreted in the light of the legal context. The interpretation of the data is further complicated by the fact that two or even more pension-relevant activities can occur at the same time. For example, a mother of a child younger than 3 years works and pays social security contributions. In this situation, the mother accumulates pension entitlements through employment and childcare credits. In the data, these pension-relevant activities are called social employment situations (File SES). A total of thirteen social employment situations are distinguished in the data, which are listed in Table 1: Table 1 Social Employment Situations in the SUF VVL 2004 Source: See Stegmann (2006, 2007). Schmollers Jahrbuch 128 (2008) 4 7These peculiarities of the data need to be taken into consideration for fertility analyses. OPEN ACCESS | Licensed under | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.128.4.647 | Generated on 2023-01-16 13:35:36