About the Survey (Memorandum)

General Information
The Goal and Methods of the Survey
Ownership and Control
Identification of an Establishment Size
Dataset
Large Business Groups
Key Findings
Additional Comments

General Information

In surveying ownership concentration in the Russian economy as a part of the work on the Country Economic Memorandum for the Russian Federation (PDF, 1.18 M) , the authors sought the opinions of sector analysts, investment banks, insurance companies and other experts knowledgeable of the actual control chains in the Russian corporate sector. It therefore deserves to be noted that the survey entails no claim of accurate knowledge of the distribution of ownership rights in any legal sense. Instead, it documents, to the best knowledge of the market, the true distribution of ownership rights in Russia in the summer of 2003.

The Goal and Methods of the Survey

The purpose of the survey was to uncover the true degree of concentration of enterprise control in Russia. Specifically, it was important to identify the ultimate owners of the controlling shares in the selected companies – or, to be more precise, the controlling parties – and the degree of their control over portions of the economy. The survey was carried out in three stages.

First, sub-sectors of the economy were selected for inclusion. The primary criterion was size, the idea being that the survey should cover as large a portion of the economy as possible. In addition, several sectors and sub-sectors were selected in consideration of their strategic importance (e.g. banks and mass media). It is noteworthy that selections were made irrespective of the degree of ownership concentration and control of firms within a sub-sector, i.e. we sought to avoid a sample bias towards highly concentrated sub-sectors.

During the second stage the initial sample of firms was drawn. The survey started at the level of “establishments,” the individual physical production units, since ownership structures, which actually demarcate firms, often needed to be brought to light. And again, the selection criterion was size: the largest firms in each sub-sector were identified for the sample. This approach was not based on an assumption that big firms are more likely to be owned by individuals with significant assets on a national scale: this is not known a priori any more than it is known whether large sub-sectors are more likely to be controlled by nationally significant owners (e.g. via large firms). The point of choosing large firms was simply to maximize coverage of each sector by employment and sales criteria. Nevertheless, the procedure is tantamount to assuming that nationally significant owners only control large firms. It thus places a lower-bound limit on the degree of concentration that will be discovered – if the largest business groups control small firms that are not included in the sample, the actual concentration of control in the economy is higher than what follows from our conclusions.

However, this approach threatened to seriously underestimate control over sales by small trading companies. In particular, transfer pricing, usually carried out during the period under consideration through the help of smaller trading companies, plays an important part in measuring the true (i.e. consolidated) cash flows of many natural resource enterprises. To arrive at a more realistic ranking of Russia's biggest business groups, therefore, we had to adjust the originally compiled dataset to take into account various profit centers and trading companies that appeared too small to be included into our original sample, yet turned out to be crucially important for obtaining accurate aggregated data for several of large business groups.

In the third stage, fieldwork was initiated. Investigators set out to identify (a) the main controlling owners of each firm and the portion of the firm they owned, and (b) any subsidiaries owned by the firms. This in turn generated new sets of firms to be investigated – owner-firms and subsidiaries. A chain would stop upwards when an “ultimate owner” or “controlling party” was identified, and would stop downwards when a firm owned no further subsidiaries. An ultimate controlling party could be an individual, a business alliance, or a corporate unit of one the following categories:

• Federal government
• Regional government
• Foreign structures and individuals
• Large Russian private owners
• Other Russian private owners

The Federal Government category includes all workers of the federal government and its organizations. An assumption was made in this case that management of all state establishments including but not limited to Gazprom, Rosneft, RAO UES, Transneft, MPS, Aeroflot and FGUP “Sevmash” is fully controlled by the federal government, and that civil servants functionally not related to the commercial enterprises act exclusively as representatives of the federal government and not as private owners

Ownership and Control

Given the problems of corporate governance in Russia, we suggested differentiating between the degree of ownership and control. Respondents had difficulty identifying the exact degree of formal ownership in some cases, because of the non-transparent nature of Russian ownership arrangements, whereas the degree of control was much more observable. Moreover, for the purposes of this survey, it is the concentration of control rather than formal ownership that matters. Nonetheless, while the questionnaire included questions on the degree of control, respondents were also asked to indicate whether that control was based on (dis)-proportional degree of ownership, or other considerations.

Identification of an Establishment Size

Another problem was determining, given the available data, the appropriate measure to use in evaluating firm size. In general, the economy-wide sales data necessary to put our sample in an appropriate context (i.e. the economy-wide data using the same sector definitions and coverage as our sample sub-sectors) can be identified only for industry – not for the newly emerging service sectors. Although sales data was acquired for most of the non-industrial firms in the sample, there is generally no data at the economy-wide level to compare them with. Economy-wide employment, on the other hand, is available in comparable sub-sectors from the Labor Force Survey (LFS). LFS data covers total employment well because it also covers small firms. Given these restrictions, we used employment data for the entire sample of companies in different sectors (with the exception of banks) and sales data as an output measure only for companies in sub-sectors of industry. The size of banks was measured by their assets.

Dataset

Gathering information resulted in a unique dataset that gives a comprehensive idea about ownership / control distribution in the Russian economy. The database contains information on more than 2.5 thousand firms, banks, individuals, business groups, and government entities. Forty five sub-sectors were chosen on the basis of playing an important role in the Russian economy, and because they can be easily identified with product markets. The sample includes the largest enterprises in these sub-sectors and, on that basis, we have been able to identify a non-trivial degree of control for 1700 firms, which employ 4.3 million people. Total sales of sample companies in industrial sub-sectors are 5.2 trillion rubles (as mentioned above, we do not use sales data for other sectors). According to comprehensive Goskomstat 2001 data for Russian industry (which includes more than 150,000 firms), our sample covers 22 percent of industrial employment and two-thirds of industrial output.

The average firm in the sample employs about 2527 workers, and has annual sales of USD 105 million. On average, we account for 88 percent of control if weighted by employment, and 92 percent of control if weighted by sales. Information on ownership and control is usually as of June 1st, 2003.

Sectoral and Sample Coverage

 Large Business Groups

To generate the list of owners which are significant on a national scale, the largest private owners in the sample were identified on the basis of firm size measured by annual sales and employment. Sales and employment were attributed to owners in proportion to their share of control. On this basis, all private domestic owners or groups of owners who control either R20 billion (USD 700 million, or 0.4 percent of the total sample) in sales or 20,000 employees (0.5 percent of total employment captured by the sample) were identified as among the largest business groups.    

These cutoffs resulted in a list ranking 22 major groups of private domestic owners who all broadly conformed to the conventional perception of a big business: the 22 include all large private owners who (i) have major stakes in firms in multiple regions; (ii) are discussed and visible in the press as Russia’s most influential and/ or wealthy asset holders, or as active lobbyists.

Key Findings

Twenty-two business groups control 45 percent of sales and 47 percent of employment in the sample: by both indicators this is more then all the rest of private owners put together. This raises the question how representative the sample is, i.e. how well its results apply to the economy as a whole. In terms of employment, we would expect that firms in the covered sectors that were not chosen for inclusion in the dataset, are too small to be part of the largest business groups. As to the (sub)sectors not covered by the dataset, there is no a priori reason why the sample should not be representative – recall, that sectors for inclusion in the database where not chosen by their expected degree of ownership concentration.

With an assumption that category “Other owners” includes firms in respect of which it was not possible to define a share controlled by an owner, and firms not selected to the sample, the bottom threshold of employment and sales shares controlled by various groups of owners may be obtained. Hence, 22 largest owners control at least 20 percent of employment and 39 percent of sales of the total economy.

Concentration of control by financial-industrial groups (FGIs) is highest by far in industry and is also high in banking. In the 32 industrial sub-sectors, FIGs control 38.8 percent of sales, whereas combined federal and regional governments control 24.5 percent (most of which is federal), and the foreign share is 5 percent. Remaining owners account for the balance of 31.5 percent.

Concentration of control is lower in terms of employment: the large owners and their groups control 20 percent of employment in the industrial sectors covered by our sample. The shares of employment controlled by the federal government, and to a lesser extent, by regional government and foreign owners, are also smaller than their shares as measured by sales.

Additional Comments

1) Although the ownership and control data refer primarily to mid-2003, the data on sales, employment and assets refer to 2001 and 2002, since more recent data for the majority of firms in the database are not yet available. Performance measures therefore had to rely on sales and employment data from these years: in other words, we need to assume that changes in control between 2001, 2002 and 2003 have been minor. Given data availability, this is an inevitable limitation (but as the most active phase of privatization and consolidation was concluded before these dates, it is perhaps not too distorting an assumption to work with).

2) The concentration results for sales use 5-digit sector aggregates from the Russian Registry of Industrial Enterprises, a census of medium-sized and large firms, which misses out sales by small industrial firms (approximately 20 percent of industrial sales on aggregate). The result, ceteris paribus, is a modest overestimate of concentration, but only when using sales data. Employment concentration results are based on data from the Goskomstat Labor Force Survey, which covers all economic activity (including unregistered employment). The omission of ownership stakes by large owners in small firms that may have, as a result of the sample’s design, escaped consideration implies a modest underestimation of concentration, independently of whether measured by sales or employment.