Beijing – China - 18 – 22 October 2004
Session No6Paper No
Raoul Depoutot, INSEE, PARIS,
25 years of Enterprise Group statistics in France
Experienced gained and current development
1 PRODUCTION OF ENTERPRISE GROUPS STATISTICS
1.1Financial links and delineation of Enterprise Groups:
In the beginning of the 80s INSEE started to conduct a survey on capital links between corporations. A group of economists from inside and from outside INSEE had emphasized the role of enterprise groups (EG hereafter) in the economy, and the need to have reference statistics on them. Only private databases of consolidated accounts were providing information on the biggest EGs. Accounting rules on consolidation were not sufficiently uniform to be suitable for statistics. Therefore INSEE had to develop its own definition of an EG, and designed a survey to identify these groups. Statisticians have defined the statistical group as a group of corporations dependent on the same parent company, giving a precise definition of the terms “dependent” and “parent company.” The parent company, called “group head,” is a corporation which is neither directly nor indirectly “controlled” by any other company and which itself controls at least one other company. The “core” of the group is the group head and the companies it controls directly or indirectly. “Control” is the ability of the group head to define the strategic orientations of the subsidiary. This control derives from the ownership of the equity capital, through the exercise of the voting rights which allow appointment of directorial bodies.
Such a definition of control is a matter of “all or nothing”: one company either controls another (if it owns directly or indirectly the absolute majority of voting rights) or it does not. The statistician will therefore define and measure the “control rate.”
The financial links (LIFI) surveyis an annual mandatory statistical survey of private corporations. It has existed since 1980. The enterprises surveyed meet at least one of the following criteria: total amount of equity interest greater than 1.2 million €, personnel in excess of 500 employees, turnover in excess of 30 million euros.
In 2003, about 24 000 corporations were surveyed. In addition INSEE is in charge of updating the Administrative Register of Enterprises Under Majority Control of the State.
Since 2002, INSEE has completed the scope of the survey with another administrative source: the information published by French corporations to the commercial courts, which encompasses not only annual accounts but as well the names of shareholders[1]. This source provides the financial links data for corporations below the thresholds of the LIFI survey[2].
The collected information consists of:
- Identification characteristics of the corporation (administrative ID number, corporation name, address, activity, legal form, number of employees, total value of majority and minority interests).
- shareholdersof the canvassed corporation: corporate name, address (at least nationality), and percentage of voting rights held by each shareholder (only for legal persons).
- composition of the shares’ portfolio: for each subsidiary or affiliate, the corporation name, address, administrative ID number, activity, percentage of capital and voting rights held, and value of the interests reported[3] in the accounts of the canvassed corporation.
The sampling list is derived from a statistical database of the accounts of all corporations registered in France, that integrates copies of the corporate tax reports. An algorithm delineates from the financial links the composition of each EG, its group head, and for each subsidiary its rank (i.e. one for direct subsidiaries of the group head, n+1 for subsidiaries of a corporation that is itself of rank n).
One of the most delicate points concerning statistics on EGs has to do with the territory in which their subsidiaries are located, because it goes beyond the national borders. This is the case when the ownership chain has several links beyond the borders: the first link is collected via LIFI, but is more difficult with lower (respectively. higher) ranked corporations for foreign EGs (respectively. French EGs).
In order to make up for this weakness of information, since 1994 INSEE has been making use of data on foreign investments in France from the administrative files of the Treasury Office[4]. This gives information not only on the direct investor, but also on the country of origin of the equity capital, that is, the country of the parent corporation of the investor group. Other available information is used as well (information published by companies). For more on the production of EG statistics, see 1.
1.2 Economic statistics on Enterprise Groups
INSEE has developed an integration system for business statistics that provides a census, based on administrative data, of financial reports for all corporations. In particular, for each corporation an income statement, a balance sheet and other sets of information are available. These accounts are non-consolidated ones. Information drawn from the annual structural business surveys adds in particular the main activity code, a breakdown of turnover by elementary activity, information on mergers and splitters of corporations, information on the workforce.
By merging the file of EGs components and the file of corporation accounts, we get for each EG economic information on all subsidiaries under its control (including the group head as well). A simple way to get estimates of the economic importance of an EG is to add up all quantitative characteristics, in order to approximate roughly the full consolidation method of the consolidated accounts. Doing that, we do not cancel out intra flows neither intra claims neither intra liabilities. Concepts that are “netted” by definition are however not impacted by this limitation. This is the case for value added, operating profit, exports and imports, employment, salaries paid, investment, tangible assets… On the contrary, some concepts are very sensitive to this kind of measurement: turnover, production, purchases of goods and services for instance. In many EGs, the internal flows of goods or services amount to an important share of the consolidated sales or purchases. This is the case as well for financial debts, since the group head or a few corporations in each EG often manage the financial resources of the whole group (see 7). For the financial surplus of the group, things become very complicated, due to the payment of dividends (see 7). The evaluation of intangible assets raises the very complex issue of the assessment of goodwill, which is currently too complex to conduct without additional information.
Although consolidated accounts are quite appreciated by financial analysts, they are not appropriate for a direct use by statisticians. The main reason is that most economic activity realised by EGs is obtained by EGs which have as well activity abroad, either because there are foreign controlled EGs or because national EGs have subsidiaries abroad. This fact is quite dominating: Of the 7.3 million employees of EGs in France, only 1.6 work in EGs that are 100% resident. There are as well other limitations attached to consolidated accounts for the statisticians: they incorporate partially controlled corporations, which could lead to double counting.
The aggregation of these corporate accounts has been simulated by the Company Monitoring Centre of the Banque de France (CGP, 2002). The comparison of both methods for 100% resident manufacturing groups (representing 145 EGs and 46 000 employees) gives the following table:
Excerpt of aggregate profit and loss account
(in million €)
Consolidated accounts / Corporate financial accounts145 Groups / % of Turnover / 554 Corporate balance sheets / % of Turnover
Turnover / 12,673 / 14,350
Value added / 4,649 / 36.7 / 4,826 / 33.6
Gross operating surplus / 738 / 5.8 / 863 / 6.0
Gross operating income / 708 / 5.6 / 938 / 6.5
Net current income before tax / 265 / 2.1 / 537 / 3.7
A comparison of sub-consolidated accounts limited to France and an aggregation such as described above for the 40 biggest listed EGs in France between 1997 and 2000 has shown that consolidated value added and operating profit are reasonably well approximated by the aggregation, whereas aggregated turnover and aggregated purchases lead to a huge overestimation of consolidated figures (intra flows represent sometimes 100% of the consolidated figures).
A main activity code for each EG can as well be derived from the activity code of each subsidiary. Different methods can be envisaged (depending if the breakdown of turnover for each corporation is used or only its main activity). Intra-sales are again a source of problems (they should not be taken into account for the assessment of the main activity). As a whole, the experience is that the appropriate level of classification for main activity in EGs is an intermediate level of about 40 “activity groups”.
Such assessment of the economic size of EGs can lead to two types of statistics:
-statistics on the population of EGs: their economic weight can be broken down by industry, by size class, by nationality and by type (private - listed or non listed -vs. state controlled).
-statistics on the whole population of businesses can be produced, using as reference statistical unit not the corporation (legal unit) but the “autonomous unit”[5], that is either the independent enterprise (corporation that does not belong to any EG) or the EG.
1.3: some examples:
Below standard statistics are provided based on the definition of EGs derived from the LIFI survey only. Therefore the real share of EGs is slightly underestimated.
Note: corporations in the manufacturing, trade and services only. Education and Health excluded.
Development of the foreign controlled economy
2- NEW INSIGHTS INTO THE FRENCH ECONOMY:
2.1: a reviewed analysis by size
The French economy has 26 000 EGs, 90 of then having more than 10 000 employees each. The great majority is of limited size (25 000 EGs have less than 500 employees). However, the 2 million independent corporations represent only about 42% of employment and 30% of value added (independent corporations are of lower size, have a lower labour productivity and hire less qualified employees). Symmetrically nearly all corporations with more than 500 employees belong to an EG.
Breaking down the population of businesses based on the corporation or on the “autonomous unit” leads to different results.
Total employment
Size of unit:Employment band / Stat. Unit: corporation / Stat. Unit: autonomous unit
(millions) / % / (millions) / %
Less than 50 / 6.0 / 44 / 5.3 / 39
50< < 500 / 3.3 / 24 / 2.7 / 20
More than 500 / 4.2 / 31 / 5.5 / 40
2.2: analysis by industry
Using the main activity of the EG as the breakdown criterion instead of the main activity at the corporation level modifies again the usual representation of the economy. In particular, some specific service activities tend to be reclassified in manufacturing. The most visible effect is of course the economic sector of holding companies, the majority of which are integrated in manufacturing EGs. The difference for characteristics like “net assets” in the balance sheet is worth noticing. Whereas the total net assets for all corporations amounts to about 1500 billion €, 370 billion are held by corporations classified in the industry of “holding companies”. When broken down according to the industry of the EGs, we get the following table:
EG industry / Total net assetsManufacturing / 290
Trades / 49
Services / 29
All / 370
Unit: billion €
2.3: analysis of concentration
Using autonomous units as analysis unit for the concentration[6] -, differs as well from the results of such an analysis that would be based only on corporations. The economic concentration[7] can be an intermediate concept, for which the reference statistical unit is not the EG, but the subset of corporations of an EG made of all corporations classified in the considered industry.
The highest sensitivity to the choice of unit for the analysis of economic concentration is observed in the “Sewage and refuse disposal industry”, where the weight of the four biggest autonomous units (EGs) is 61% of the total industry, compared with only 11% for the biggest four corporations.
2.4: some comments on statistical units and organisation of EGs
Our first experiments in profiling statistical enterprises, i.e. grouping corporations of the same EG when they have no organisational autonomy, have conveyed us the impression that enterprises and sub-groups are very similar. Such sub-groups are considered as well by EGs as the relevant level to analyse their economic activity - perhaps except for pure financial management. Some analysis has shown as well that the financial function is not limited to the group head, but involves several corporations in many EGs. Picart (7) has proposed the concept of “control centre” of an EG, as opposed to its “production basis”.
2.5: globalisation
EGs are at the core of the internationalisation of companies: a perfect knowledge of the international structure of EGs would lead to a high level of statistical information on globalisation. Currently, the analysis of resident foreign corporations is very developed, since both the list of corporations under foreign control and their accounts are available at INSEE. This leads us to produce long series of inward FATS. Symmetrically, the knowledge of subsidiaries abroad is under the responsibility of another directorate of the Ministry of Finance (DREE), which conducts every second year a census of French owned corporations abroad. The available economic information is however currently limited[8] (only employment and turnover are available). INSEE is just about to complete the production of the first complete EG file, i.e. a file that fully incorporates three sources: the LIFI survey, the complementary commercial court information, and the information from the DREE census. The main preliminary results are given in the table below.
Total employment by category of businesses (millions) in 2000
In France / AbroadIndependent corporations / 5.4
Enterprise groups: / 7.3
Foreign / 2.2
French ones without foreign subsidiary / 1.6
French ones with foreign subsidiaries / 3.5 / 3.5
2.6: economic analyses
Examples of analyses on various topics conducted during the last years may highlight the importance of statistical information on EGs.
2.6.1: Flows of dividends paid by businesses:
The growth of dividends paid by corporations from 1991 to 1999 amounts to 30 billions. This growth can be decomposed as follows, taking into account the structure of EGs: growth of dividends outside groups +3, growth of operating profit +4.5, growth of dividends paid to ultimate shareholders +3.5, growth of the share of EGs in the economy +7.5, growth due to Intra-group restructuring +11.5. The usual explanation given by economists of the growth was based on an increased pressure of ultimate shareholders on companies. This decomposition shows that this component is quite limited. On this issue, see 6.
2.6.2: Indebtedness:
In order to assess the level of indebtedness at the end of the financial bubble of 2000, INSEE analysed the indebtedness of businesses by type of autonomous unit. The result was that the most indebted units where EGs, and particularly the biggest ones that where the most committed in international growth. This is shown in the graphs below.
The survey distinguishes two categories: EGs listed on the stock market in the reference basket of the financial SBF120 index (biggest 120 capitalisations of the market) and other listed EGs (hence smaller).
The evolution of internationalisation is in line with the ratio of the wage bill for the French part and the wage bill at the world level between 1998 and 2000. The evolution of indebtedness is measured by the growth rate of the consolidated debt between 1998 and 2000. For each of these two measures of change, the groups were sorted in ascending order and distributed into three classes of equal size. All listed non-financial companies for which this information is available over three years were included in the field. To learn more about the analysis, see 7.
2.6.3: demography of businesses
By combining the corporation dimension and the EG dimension, Picart (8) has just shown interesting messages drawn from facts: first, that there has been nearly no real “success story” in the French economy compared with the US economy (i.e. there is no EG in the biggest ones that was not present some 20 years ago as such or as the start of further restructuring). Second, the global dynamics is that the share in the economy of medium sized autonomous units has diminished, whereas bigger EGs have maintained their share by the purchase of dynamic corporations they tend to support and by the sale of less dynamic corporations that were part of these EGs. This is to compare with the growth abroad of bigger EGs, in particular those listed in the CAC40 index of the stock market (see 4).
2.6.4: mobility of staff in EGs
Duhautois and Delarre (5) have merged the Financial Links survey (LIFI) with the panel of annual payroll declarations of individual employees by corporation. They have quantified the mobility of EG employees in France from 1991 to 1999. The analysis shows that this internal labour market of EGs is made up mainly of geographically close corporations and structurally similar corporations in the EG. This finding holds regardless of the main activity of business and EG size. However, the role of size is non-linear (it can rather be represented by a U curve).
2.6.5: analysis of financing of businesses:
In 2001, the French Planning Agency[9] (CGP) launched an investigation into the financing of the French economy, in particular on the growing influence of stock markets as in many industrialised countries. The policy issue behind it was to consider if the growing role of own capital in the financing of companies should still be supported as a mean to achieve faster growth. Among other things, the working group investigated the impact of the then famous “standard” for financial return of listed companies (i.e. a Return On Equity of about 15%) and its sustainability. The group consisted of macroeconomists, auditors, financial analysts, national accountants and public statisticians. Economists considered various statistics on profitability, and the interpretation of these statistics was not an easy task. Mainly three types of statistics were available: statistics on the stock market, which were then showing very high profitability rates, statistics drawn from the national accounts, which were showing a much more limited profitability and statistics drawn from business statistics, which were providing a profitability level differing from the two other statistics.
Some economists were not particularly dissatisfied with the situation and could cope with it, and provided economic explanations they felt compatible with it. Others, and especially those working or having worked in the economic directorate of INSEE[10], were more disturbed and asked for an in-depth investigation of the concepts and methodology defining these statistics.