The Establishment File of the German Social Insurance Statistics

Fritsch, Michael; Brixy, Udo Article The Establishment File of the German Social Insurance Statistics Schmollers Jahrbuch – Zeitschrift für Wirtschaftsund Sozialwissenschaften. Journal of Applied Social Science Studies Provided in Cooperation with: Duncker & Humblot, Berlin Suggested Citation: Fritsch, Michael; Brixy, Udo (2004) : The Establishment File of the German Social Insurance Statistics, Schmollers Jahrbuch – Zeitschrift für Wirtschaftsund Sozialwissenschaften. Journal of Applied Social Science Studies, ISSN 1865-5742, Duncker & Humblot, Berlin, Vol. 124, Iss. 1, pp. 183-190, https://doi.org/10.3790/schm.124.1.183 This Version is available at: https://hdl.handle.net/10419/292081 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/ Schmollers Jahrbuch 124 (2004), 183-190 Duncker & Humblot, Berlin The Establishment File of the German Social Insurance Statistics By Michael Fritsch and Udo Brixy 1. The German Social Insurance Statistics The German Social Insurance Statistics (,,Statistik der sozialversicherungs pflichtig Beschäftigten", SIS) was introduced in West Germany in 1973. In eastem Germany, it became compulsory from January 1991 on, shortly after unification. The SIS requires all public and private employers to report certain information about every employee who is subject to obligatory social insur ance, i.e. health and unemployment insurance along with pension funds. The information collected can be transformed into an establishment file that pro vides longitudinal information about the establishments and their employees. In this contribution we give an overview of the content of the SIS establish ment file as well as of the possibilities and the problems involved in using this data for scientific analyses. We first describe the coverage and information contained in the SIS (section 2) and give a brief outline of the process by which this information is raised and prepared. Section 3 deals with data errors and reliability problems, particularly the definition of an establishment and the identification of start-ups and closures. Section 4 depicts the main topics of studies based on the data and describes the current possibilities of data access. 2. Coverage and informational content The German SIS is derived from information that every employer is obliged to report about those employees who have to contribute to the social insurance system. The statistics therefore contain only data about this group of persons. Civil servants, army personnel, self-employed and unemployed persons are not included. Due to some changes of the criteria for being obliged to contri bute to the social insurance scheme there was some variation in the definition of the observation unit over time. In manufacturing, the SIS currently covers more than 90 percent of all employed persons. In the service sector, this share is about 80 percent. lt is relatively low in agriculture (less than 24 percent) and in the public sector (about 50 percent). The employers report the informaSchmollers Jahrbuch 124 (2004) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.124.1.183 | Generated on 2023-04-04 12:32:59 184 Michael Fritsch and Udo Brixy tion to the responsible health insurance company, which passes it on to the statutory pension insurance. The pension insurance transfers it to the Federal Employment Service (Bundesanstalt für Arbeit) where the data are collected and prepared. Each of these institutions may check the accuracy of the data and correct errors. For every employee who is obliged to contribute to the social insurance system the SIS contains the following information: • social insurance number; • gender and date of birth; • a number for the establishment where the person is employed; • code number of the municipality (Gemeindkennziffer) in which the estab lishment is located; • industry affiliation (since 1998 according to the five-digit NACE classifica tion system); • profession according to a classification of 334 categories; • employment status ( e.g. apprentice, unskilled, skilled, master craftsperson; füll time employees only); • the educational level attained and vocational training concluded; • nationality and • gross income (truncated at the income limit at which employees are free to opt out of the statutory health insurance and choose private health insur ance ). For those establishments that belong to a firm with several establishments, no information about the respective firm is contained in the data. 3. Different forms of data preparation The most frequently used form of the SIS to date is the quarterly end of period files (Vierteljahres-Stichtagsdateien). These files contain information about every employed person that was subject to compulsory social insurance on the last day of each quarter, i.e. as of March 31, June 30, September 30 and December 31. The quarterly end of period files are available about nine months after the respective date. lt has become common practice in scientific analysis of these data to prefer the use of the quarterly file of June 30 for two reasons. First, there is no indication of any specific bias on that date. And second, information relating to the middle of a year appears to be particularly representative of that year. Another form of the SIS is the history file (Histo rikdatei), which contains all of the information gathered about a certain emSchmollers Jahrbuch 124 (2004) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.124.1.183 | Generated on 2023-04-04 12:32:59 The Establishment File of the German SIS 185 ployee since the introduction of the statistical system. lt makes it possible to follow an employee's work history. Because the history file includes even information that is reported very late and also accounts for late error correc tions, it is the most complete and reliable version of the data. Although the original unit of observation of the SIS is the employee, the data can be transformed into an establishment file by means of the establish ment number that is reported. In this procedure, all employees with the same establishment number are allocated to this number. The current version of the establishment file is generated from the quarterly end of period files that relate to June 30 of each year. These data comprise information about all establish ments in the Federal Republic of Germany that report employees who were obliged to contribute to the social insurance system on at least one of these dates. Because the SIS does not contain any information about the respective firm, the establishment file can not be further aggregated to that level. The current West German file goes back as far as 1981 and does not cover the earlier years because the respective quarterly files no longer exist. Currently, a new version of the establishment file is being generated on the basis of the history file. This new data set will provide information from the introduction of the SIS in 19731 onwards and it will be more comprehensive and reliable than the establishment file based on the quarterly end of period data. 4. Data errors and reliability problems 4.1 The definition of an establishment The definition of what constitutes an establishment is of crucial importance for the usefulness of the establishment file. Generally, the establishment iden tifies a certain location, i.e. a plant or a workplace. There are, however, two main deviations from this definition. First, if a firm has several plants in one municipality which are affiliated to the same industry, the employees in these plants may be subsumed under the same establishment number. And second, a certain establishment may be split into several establishment numbers if this seems appropriate to the management for any reason. This could be the case particularly in relatively large establishments with different lines of operation. The fact that the establishment number of the SIS does not relate to a certain plant or location can be a problem particularly in the public sector, in non profit organizations and in industries where larger firms tend to have many small branch offices such as in the retail trade, banking and the insurance industry. However, such incorrect definitions of the establishment are rela1 However, due to diverse errors during the implementation of the SIS the quality of the data gathered in the first two years must be regarded as rather inadequate. Schrnollers Jahrbuch 124 (2004) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.124.1.183 | Generated on 2023-04-04 12:32:59 186 Michael Fritsch and Udo Brixy tively rare (well below ten percent of all cases) and large deviations of the number of employees as recorded in the SIS from the 'true' number are an exception (cf. Brixy and Fritsch 2002, 74 f.). 4.2 Identification of start-ups and closures Under the assumption that new establishment-numbers in the German Em ployment Statistics may be regarded as indicating newly founded establish ments and that disappearing numbers represent closures, the establishment file of the SIS provides the opportunity of analyzing establishment dynamics in the economy. In fact, the start-up of a new establishment with employees that have to contribute to compulsory social insurance will usually2 lead to the inclusion of a new establishment number. And in the case of a closure, the number will normally disappear. However, there are other possible reasons for the appearance and disappearance of an establishrnent number in the statistics, such as • the redefinition of an existing establishrnent, which can lead to either the cancellation of an established number and the introduction of a new one, • ownership change, and • in some rare cases the alteration of a firm's legal form. Because establishments may exist for some considerable time before hiring an employee, start-ups may appear in the SIS 'too late'. Correspondingly, clo sures may disappear from the SIS prematurely if the establishment continues to exist without any employees. One main problem in identifying start-ups and closures is establishments with 'perforated' employment, where information is interrupted for some time. In this case, it is unclear whether in the years with rnissing information the respective establishrnent • has provided the data too late to be included in the quarterly end of period files which are currently the basis for the establishrnent file, • had existed without an employee, or • had been closed down and the entrepreneur has set up a new business after some time.3 2 lt will not lead to a new number if the new establishment belongs to a firm with an existing establishment of the same industry in that municipality and if the parent firm wants to subsume these establishments under one number. 3 Tue establishment number is linked to the person of the owner. If an entrepreneur has gone out of business and then starts a new firm after some time, the old establish ment number is used for the new business. Schmollers Jahrbuch 124 (2004) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.124.1.183 | Generated on 2023-04-04 12:32:59 The Establishment File of the German SIS 187 Because establishments with a perforated number tend to be rather small4, one can assume that in many cases they have existed without an employee during the interruption.5 In the current version of the establishment file, the length of the time period for which there is no information about an establish ment in the data is used to separate continuing establishments from start-ups and closures.6 Establishments with an interruption of less than three years are assumed to have continued without an employee during that time. If an inter ruption has lasted longer than three years, the disappearance of the establish ment number is classified as a closure and re-emergence as a start-up. The range of possible rnisidentification associated with this criterion can be esti mated in a rather simple way. According to the criterion, about 87 000 estab lishments that reappeared in the 1981-97 period are classified start-ups. If all of these establishments had continued business during the time of interruption, the number of start-ups would be overestimated by about four percent. If about half of these establishments had continued, the rnistake would amount to about two percent. There are a quite a lot of new establishment-numbers in the data that cover an untypically large number of employees. One might suspect these new num bers to be related to the reorganization of an existing firm or to the establish ment of a larger branch plant. In order to single these cases out, the number of employees is used as a second criterion for identifying start-ups. New estab lishment-numbers are not classified as start-ups if more than 20 employees are reported in the first year. Therefore, a new establishment-number is classified as a start-up if it was not included in the establishment file during the three preceding years and does not report more than 20 employees on the date it enters the statistics. Because no additional size criterion is applied for the identification of a closure, classification as a closure is somewhat less restric tive than classification as a start-up. As a consequence, these definitions lead to an underestimation of net-entry, i.e. the number of start-ups minus the num ber of closures. However, comparisons with the data of a yearly interview survey of some 4,000 firms which is based on a random sample of the SIS establishment file (Establishment Panel of the German Institute for Employment Research, IAB) 4 94.6 percent of all West German establishments that were characterized by perfo rated information in the 1981 -97 period had no more than three employees in the year before the interruption occurred. 77.8 percent had only one employee. 96.1 percent of these establishments reported no more than three employees when they re-appeared in the statistics. 80.3 percent had only one employee. s Interrupted information due to late notification does not occur if the history file is used for generating the establishment data. 6 Taking the West-German data, we find that between 1981 and 1997 about 60% of all interruptions were just for one year. Only 14.5% of all interruptions related to a period longer than three consecutive years. Schrnollers Jahrbuch 124 (2004) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.124.1.183 | Generated on 2023-04-04 12:32:59 188 Michael Fritsch and Udo Brixy showed that these relatively crude assumptions lead to quite satisfactory re sults (Brixy 1999). Furthermore, comparisons of different data bases for iden tifying start-ups in Germany have revealed that errors and mis-specifications of start-ups were in no way more relevant in the SIS than in the other data sources (Fritsch et al. 2002). Moreover, the numbers of start-ups by year, in dustry and region that were identified from the different data sources tended to be quite closely correlated. So even if the information about start-ups gener ated from the establishment file of the SIS can not be regarded as the true number of newly founded establishments, it represents quite a good indicator of the actual number. There is currently no other data base available in Ger many from which closures can be identified as comprehensively and reliably as from the establishment file of the SIS (cf. Fritsch et al. 2002). Tue level of errors and mis-specifications in identifying closures should be in about the same range as for start-ups, so the data can be regarded as an appropriate indicator of the 'true' number of closures. The higher data quality of the new version of the establishment file that is based on the history file of the SIS can be expected to perrnit a considerable reduction of errors in identifying clo sures and start-ups. 5. Main topics of studies based on the data and data access Empirical studies based on the establishment file of the German SIS have focused on start-up processes, particularly regional variations of the number of newly founded firms7, the effects of new firm formation on econornic de velopment8, the performance of establishment cohorts9 and job-tumover10. Direct access to the SIS is limited to the employees of the Federal Em ployment Service and the associated IAB. Currently, some aggregate infor mation ( e.g. the number of employees with certain qualifications in different industries or regions) are published on a regular basis. In the future, informa tion from the establishment file of the SIS will be available from a research data centre that is planned to be set up at the IAB. An anonymized one percent random sample drawn from the history file that provides information on the number of employees and also about the respective establishments can be made available for scientific analysis (cf. Bender, Raas and Klose 2000). Information about the number of start-ups and closures per year, in dustry and region are provided by the Chair of Econornic Policy at the Tech7 Audretsch/Fritsch (1994, 1999), Brixy (1999a), Brixy/Grotz (2002), Fritsch (1992), Fritsch/Falck (2003), Lehmann, Mössinger (1996), Lehmann (1996a), s Audretsch/Fritsch (2002), Fritsch (1996, 1997). 9 Boeri/Cramer (1991), Brixy (1999a und b), Brixy/Kohaut (1999), Brixy/Grotz (2004). 10 Cramer/Koller (1988) Lehmann (1996b), Brixy (1999a). Schmollers Jahrbuch 124 (2004) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.124.1.183 | Generated on 2023-04-04 12:32:59 The Establishment File of the German SIS 189 nical University of Freiberg and can be downloaded from the intemet under http:// www.wiwi.tu-freiberg.de/ wipol/ gründungsatlas / gründungs_indika toren. References Audretsch, D. B./ Fritsch, M. (1994): The Geography of Firm Births in Germany, Regional Studies 28, 359-365. -(1999): The Industry Component of Regional New Firm Formation Processes, Review of Industrial Organization 15, 239-252. -(2002): Growth Regimes over Time and Space, Regional Studies 36, 113 -124. Bender, S. / Haas, A. / Klose, Ch. (2000): The IAB Employment Subsample 1997 - 1995, Schmollers Jahrbuch 120, 649-662. Boeri, T. / Cramer, V. (1991): Betriebliche Wachstumsprozesse -Eine statistische Ana lyse mit der Beschäftigtenstatistik 1977 -1987. in Mitteilungen aus der Arbeits marktund Berufsforschung 24, 1, 70-80. Brixy, V. (1999a): Die Rolle von Betriebsgründungen für die Arbeitsplatzdynamik. Bei träge zur Arbeitsmarktund Berufsforschung 230, Nürnberg. -(1999b): Gründe und Hintergründe des Erfolgs neuer Betriebe in Ostdeutschland, in: Wiedemann, E. et al. (eds.): Die arbeitsmarktund beschäftigungspolitischen Heraus forderungen in Ostdeutschland. Beiträge zur Arbeitsmarktund Berufsforschung 223, Nürnberg. Brixy, V./ Fritsch, M. (2002): Die Betriebsdatei der Beschäftigtenstatistik der Bun desanstalt für Arbeit, in: Michael Fritsch and Reinhold Grotz (eds.), 55-77. Brixy, V./ Gratz, R. (2002): Räumliche Differenzierungen von Betriebsgründungsinten sität und Überlebenschancen in Westdeutschland 1983-1997, in: Raumforschung und Raumordnung 60 (2), 100-122. -(2002): Räumlicher Vergleich der Datengrundlagen zum Gründungsund Still legungsgeschehen in Deutschland, in: Michael Fritsch and Reinhold Grotz (eds.), 165-198. -(2004): Entry-Rates, the Share of Surviving Business and Employment Growth: Dif ferences between Western an Eastern Germany since Reunification, in: Michael Dowling, Jürgen Schmude and Dodo zu Knyphausen-Aufsess (eds.): Advances in Interdisciplinary European Entrepreneurship Research, Münster: Lrr, 143-152. Brixy, V./ Kohaut, S. (1999): Employment Growth Determinants in New Firms in East ern Germany. Small Business Economics, 13, 155 -170. Cramer, V./ Koller, M. (1988): Gewinne und Verluste von Arbeitsplätzen in Betrieben - der „Job-Turnover"-Ansatz, in: Mitteilungen aus der Arbeitsmarktund Berufs forschung 21 (3), 361-377. Fritsch, M. (1992): Regional Differences in New Firm Formation: Evidence from West Germany, Regional Studies 25, 233-241. Schrnollers Jahrbuch 124 (2004) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.124.1.183 | Generated on 2023-04-04 12:32:59 190 Michael Fritsch and Udo Brixy -(1996): Turbulence and Growth in West-Germany: A Comparison of Evidence by Regions and lndustries, Review of Industrial Organization 11, 231 -251. -(1997): New Firms and Regional Employment Change, Small Business Economics 9, 437-448. Fritsch, M.! Falck, 0. (2003): New Firm Formation by Industry over Space and Time: A Multi-Level Analysis, Discussion Paper 322, German Institute for Economic Re search (DIW Berlin). Fritsch, M. / Gratz, R. (eds.) (2002): Das Gründungsgeschehen in Deutschland -Dar stellung und Vergleich der Datenquellen, Heidelberg: Physica. Fritsch, M. / Niese, M. (2002): Vergleich auf gesamtwirtschaftlicher und sektoraler Ebene, in: Michael Fritsch and Reinhold Grotz (eds.), 141-164. Fritsch, M. et al. (2002): Zusammenfassender Vergleich der Datenquellen zum Grün dungsgeschehen in Deutschland, in: Michael Fritsch and Reinhold Grotz (eds.), 199-214. Lehmann, V. (1996a): Dynamik und Beschäftigungsentwicklung der ostdeutschen Be triebe 1991-1995. MittAB 29 (4), 580-588. -(1996b): Flexibilität durch Job-Tumover und Betriebsgründungen.in: Flexibilität von Betrieben in Deutschland. Beiträge zur Arbeitsmarktund Berufsforschung. 200, 105-118. Lehmann, V./ Mössinger; W. (1996): Regionale Arbeitsplatzdynamik in den neuen Län dern. In: Informationen zur Raumentwicklung, H.1, 29-38. Schmollers Jahrbuch 124 (2004) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.124.1.183 | Generated on 2023-04-04 12:32:59