Survival of state-owned enterprises

Survival of state-owned enterprises

Below, please find the link to download the unique dataset compiled as part of PRIVATIZATION project. The list of firms is based on the list from the book by A. Karpiński, S. Paradysz, P Soroka, and W. Żółtkowski entitled Jak powstawały i jak upadałby zakłady przemysłowe w Polsce [On the rise and fall of manufacturing plants in Poland], published in 2013. This book presents a unique record of all manufacturing plants from the Central Planning Committee from 1988. We subsequently extended and updated these records. The extensions include the history of firms since 1988 until today with a special focus on the eventual bankruptcy, privatization and involvement of foreign investors. These data are publicly available are available in the form of SOEs-pedia, an interactive encyclopedia of the identified firms as well as a spreadsheet dataset that is ready for researchers to use. We briefly describe data collection procedure below. Further down you will find the list of variables as well as the download links. The database distributed here contains K/L ratios, contact us for original K and L levels.


Data collection procedure

There is no single comprehensive source of data about the firms (or plants) existing in Poland before the transition. Karpinski et al. (2013) provide a complete list with data on sales, employment and capital stock. Still, this source does not match the privatization registries of the Ministry of Treasury, neither does it match fully the records of SOEs held in early transition by the Ministry of Finance. The list provided by Karpinski et al. (2013) is substantially longer than the alternative sources. To inquire about the transition history of these plants, we exploited two sources of data: Internet and Court Archives.

Internet Search - Plant-level data from the Internet was collected in a threefold manner.  First, we searched for information using accessible public records. Second, we searched for any relevant articles available in Polish newspapers as well as for information on a given company's official website. Last, we used relevant keywords in Google searches to find any additional information.

[1] In the beginning of our search, we found the appropriate identification numbers for a given firm, namely the REGON, NIP, and KRS. Using any one of these three I.D.s, we were able to search for basic information about the firm (for example, whether it is still active, and if not, when it stopped functioning) by inputting them into the engine on the Central Statistical Office's website (Wyszukiwarka GUS) and, additionally, into third-party sites offering digital recordkeeping services ( and IMSiG). Information provided by these websites included the date of formation, date of liquidation, and the date of removal from the registry. Typically, a state-owned enterprise retained its I.D. when it was restructured and privatized. However, in some instances, when the newly formed private company was not a direct continuation of the former SOE, it was registered as a completely new entity. This would, for example, happen when the new investor of the firm changed the business and production profile of the company. At the same time, a privatized SOE could be recognized as a completely new firm if other parts of its capital were sold off or liquidated. Since our data is plant level, we considered a newly established firm to be a continuation of an SOE when the plant specific to our database remained functioning following the restructuring.

[2] If not all of the desired data could be obtained through information linked to the identification mentioned above, we searched for websites linked with the specific plant, firm, or industry. Most existing firms with websites feature some sort of "about us" section with relevant historical information. The detail with which the history of an existing plant was explained on a company website varied considerably, and therefore could not be taken as the sole means of identification for surviving plants. Further sources, mostly national and local newspapers, had to be used to gain more in-depth insight into the current status of some of the plants. Newspaper and media sources additionally contributed to the information on liquidation and changes in ownership. For plants that were liquidated, restructured, or privatized in the 1990s and early 2000s, we searched the digital archive of the daily newspapers and weeklies (both national and regional).

[3] When the two above methods failed to yield the information we desired, we did a basic Google search on the given plant/firm.  The specific search terms were:

  • plant name* + "prywatyzacja" OR "podział" OR "inwestor" OR "FDI" OR "zagraniczny inwestor" OR "spółka akcyjna" OR "spółka skarbu państwa" OR "akcje" OR "sprzedaż"[1]; and
  • plant name* + "upadłość" OR "likwidacja" OR "podział" OR "nierentowność" OR "sprzedaż mienia" OR "sprzedaż gruntu" OR "zamknięcie"[2]

To ensure the reliability of the information provided in the search results, we sought to find multiple sources that verified given facts about a plant. Only when a date or event was reported by two or more sources was it included in our dataset.  These sources had to be independent of each other. In that, two websites that cited the same source could not be taken as two separate sources. Thus, only primary sources were considered factual.

Please note that the name is typically not sufficient. For example, there are 12 "Zakłady Ceramiki Budowlanej" (Construction Ceramics Plant) in our sample, each in a different location, possibly also different specialization. To identify the fate of the plant, one needs to identify a specific plant, with location matching.

Court Archives. We collected data from the Regional Commercial Court in Warsaw (pol. "Sąd Okręgowy w Warszawie – Sąd Gospodarczy"). The registry of public firms (RPP- pl. Rejestr Przedsięborstw Państwowych) found in the court was used for information about the potential privatization or liquidation of state-owned enterprises.  The records available in several regional courts were queried, but the procedure was not carried out for all provincial courts because of its inefficiency. First, most of the information from the RPP is also available online in alternative sources (RPP holds the source documents, but the status has been recorded in several digital registries). Second, a large number of privatizations involved at the initial stage a transformation from a state enterprise to a treasury-owned company. Given that our interest encompasses the complete privatization process, RPP records conclude too early for many plants.  Third, RPP contains information from a firm level, whereas our data are available at the plant level.

Business Intelligence Following all of the above steps, we tried to fill in the gaps of our dataset by purchasing information from two business intelligence firms – a domestic and international one – whether they could find any information regarding the status, transformation, fate, employment or sales, of some of the plants for which we could not find information. Both of the publishers informed us they were unable to find any relevant information.

Checking the reliability of information: business Intelligence. We addressed two business intelligence firms – a domestic and international one – with the request to provide transition histories of the plants from Karpiński et al. (2013). A randomly selected subsample of both identified and unidentified plants was provided with the request. Both of the business intelligence firms were unable to enrich our database.

On the whole, among 1,641 Polish plants in the Karpinski et al. (2013) database, 56 percent are identified. Noticeably, the identification varied significantly based on sector: sectors with high capital concentration had more transparent histories, whereas plants such as food processing or clothing have more frequently disappeared without an identifiable legal trace. Moreover,  extra 25 firms have been identified partially, that is, they are clearly non-existent (and were not privatized), but we were not able to determine the time of death, rendering these observations unusable for survival analysis. In 5 cases, we were not able to identify the year of privatization, although we have identified the former SOE to be a functioning private firm.


  • Name and City is the plant/firm identification and comes from Karpiński et al. 
  • Address is the street address and is based on own research in to the plant. Once we know the address, we also know Region and Voivodeship
  • NACE rev. 1 is the NACE rev. 1 classification of the firm and comes from Karpiński, A., Paradysz, S., Soroka, P., Żółtkowski, W. Jak powstawały i jak upadałby zakłady przemysłowe w Polsce. Wydawnictwo Muza, 2013. Based on this, we also provide NACE rev. 2
  • KLRatio is the capital to labor ratio, calculated on the basis of capital and employment statistics data, in mln of 1988 Polish Zloty per person employed.
  • Year of Birth (SOE) is the year in which a given plant was established. The original data comes from Karpinski et al. However, our research revealed a number of cases where this dataset was inaccurate, we report the most accurate data obtained. 
  • Year of Birth (Historical) gives the historical data of industrial enterprise being founded, if SOE was created on already existing premises. It is based on our own research. 
  • Year of Transformation is the year in which SOE changed its legal form from a part of the state to a potentially state-owned enterprise. This was a first step towards privatization, it required due diligence. We also provide Year of death in the case of transformed firms and the Year of liquidation in the case of non-transformed firms. Consequently, we also provide a binary variable Survival, taking on the value of 1 if any part of the plant survived as a firm until 2018.

Once we identified the plant, we sought all firms that replaced it. This search material revealed the identification numbers of all plants: NIP (tax ID), KRS (registration id with the court for legal persons) and REGON (economic activity registration number). All three are reported in the data, permitting all individuals to trace subsequent histories of the plants in our data, merging with other datasets, etc. 

We also developed a wide battery of privatization measures:

  • Year of privatization, which denotes the year in which we were able to identify that majority ownership is private. Not all firms remained private, so we also developed a variable which denotes if eventually this firm is private (EventuallyPriv), which is a binary which takes on the value of 1 if a firm is still private in 2018 (or was liquidated as a private firm). Based on this variable we also define a dummy variable Priv, taking on the value of 1 if the firm was privatized.  
  • For privatized firms, we also distinguish if they were privatized to domestic or to foreign owners: directly or eventually. The dummy variable ToForeign takes on the value of 1 if the entity was directly privatized to a foreign owner. The dummy variable EventuallyForeign takes on the value of 1 if the firm became or remained predominantly foreign owned in 2018.