The OECD MAGIC (MAnufacturing Groups and Industrial Corporations) database is a confidential firm level database combining basic financial and economic data and estimates of government support at the level of each manufacturing group covered (OECD, 2025[19]). The data collected include detailed information on the ownership structure of firms, enabling the OECD to measure the scale and prevalence of government ownership in the 15 industrial sectors covered thus far in the database (Figure 1).
In collecting data on the ownership structure of companies, it becomes apparent that government ownership can vary widely across industrial firms. While some companies are fully privately owned and others 100% owned by their government, many fall in between these two extremes. Some are majority-owned by government, while others are government-invested, with government ownership accounting for less than 50%, 25%, or even 10% of their shares. In that sense, government ownership of individual firms should not be understood in a binary fashion but rather as a continuum, ranging from firms with zero government ownership to firms in which governments hold all the shares. For this reason, OECD work on industrial subsidies traditionally distinguishes between four buckets of government ownership, namely companies with:
(i) less than 10% government ownership shares;
(ii) between 10% and less than 25% government ownership shares;
(iii) between 25% and less than 50% government ownership shares;
and (iv) more than 50% government ownership shares.
Moreover, companies may be owned by different government entities serving as direct or indirect shareholders. In this respect, to measure the percentage of shares held by the government in each company covered in the database, the OECD takes account not only of direct ownership of the share capital by one single government entity1 but also of more complex ownership structures occurring horizontally through cumulative ownership (namely the accumulation of shares in the hands of different state entities), as well as indirect government ownership structures occurring vertically through a chain of ownership. In the case of direct government ownership dispersed across various government entities, be they governmental agencies, sub-central authorities or other SEs, the OECD works on the basis of the cumulative ownership resulting from the addition of shares owned by the different government entities. In the case of vertical linkages involving other SEs as owners or chains of shell entities serving as intermediaries between government bodies and the companies they own, the OECD aims to calculate the shares owned by the ultimate state entity (i.e. the ultimate state controller) proportionally to the shares it owns in the different SEs or shell entities separating it from the company.
This method aims to approach government ownership in a factual and measurable way by identifying and quantifying governments’ direct and indirect investments in firms. In differentiating between different levels of government investment, the OECD does not, however, prejudge the implications such investments may have for the effective level of state control of companies and the legal treatment that may result.2 In addition, although it represents a key element to understand the extent of government presence in, and control of, firms, government ownership of a company’s shares is not the only means through which governments can exert control over firms. Additional consideration may need to be given to the voting rights held by the government in the company, the presence of government representatives on the boards or any other equivalent management body, governmental ownership of golden shares, and other relevant factors.