22/12/2010
ESSnet on profiling large and complex MNEs: 1st SGA.
Executive summary
1 Present situation of enterprises groups and enterprises
2 From concerns …to profiling
3 The profiling options
4 A consistent system of Statistical Units
5 An example of the consequences on SBS statistics
6 Consultation of shareholders and of deciders
7The work programme for 2011/2012
1 Present situation of enterprises groups and enterprises
1.1 In the present situation, the statistical units concerning the enterprise groups are defined in the last Business Register Regulation (2008), and as such are included in the European Group Register that is being developed.
The enterprise unit is the main statistical unit of the Structural Business Statistics and is operationally largely used in this context.
1.2 The definition of the enterprise comes from the Statistical Units Regulation passed in March 1993; it is the following:
“An Enterprise is:
- the smallest combination of legal units
- that is an organisational unit
- producing goods and services (for the market)
- … with a certain degree of autonomy in decision making
- especially for the allocation of its current resources”.
While this definition seems easy to use, and notionally is not limited to national boundaries, there is some lack of clarity and certainly only limited experience in applying it. Its main strength is that it is based on the legal units which in the context of 1993 were all supposed to be domestic, are readily accessible, country by country, through administrative or fiscal data and adapted to underpin National Accounts. Its weakness is that it does not appear to be consistent with the management structure of some large and complex MNEs and is difficult for NSIs to apply consistently.
Among the characteristics defining an enterprise, “autonomy” appears essential and has found no fundamental critics, even if autonomy is supposed to be of “a certain degree”, which is all but evident, and can be limited to the “allocation of its current resources”, which may be difficult to apply when (physical) investment is concerned.
1.3 In practice, the enterprises are created nationally, by each National Statistical Institute, on its own statistical and/or legal basis, without any kind of checking the international comparability:
- if you refer to the Enterprise Group the Enterprises belong to, and whenever you would try to compare them, you would usually find that the sum of the national parts (the enterprises) differs from the global Group total or from its European part;
- the required geographical level in which the definition is supposed to apply is not specified (it is presumed that the EU regulation currently stands on a national basis but some of its criteria like “autonomy” go beyond the national borders. No criterium specifies how to assess autonomy in Europe – including or not EFTA - or at world level).
- the way the enterprises are used in data collection is not standardised and in fact non consistent between countries.
For MNEs, this means that their national parts are differently approached by the relevant national statistical institutes for the collection of similar data.
1.4 The increasing “globalisation” has heavy consequences of the following types given as examples:
- the operational structures of businesses deviate from the administrative legal structures more and more frequently, creating as said previously a deepening gap with managerial organisations and data reporting systems.
- it amplifies problems in statistics: as the intra group flows become higher and higher, the transfer flows and prices need to be recognised as such or suppressed; imports and exports change their nature, often becoming intra-group and incorporating less traded goods; the attribution of value added (VA) among countries can become internal to enterprises groups; R&D and ancillary activities are often produced in one country for the whole group etc.
1.6 In consequence, National Statistical Institutes (NSIs) consider the quality of the data on MNEs, and also all-resident enterprise groups (domestic), with a complex operating and/or financing structure as a serious issue of concern. If quality of the national economic statistics is affected, this also has a negative impact on the quality of European statistics compiled from these national statistics, resulting in concerns from Eurostat regarding quality.
In short, national as well as European statistics are rapidly loosing their accuracy and validity (correctness).
2 From concerns …to profiling
2.1 The questions implied have been studied for a long time on national and international scales.
- The NL and the UK have created new profiled units for twenty years and used them operationally ever since.
- In FR a decision to change the statistical units for a large part of the statistical system including SBS and STS has been taken some years ago and the 2011/2012 statistical programme will profile the top 100 largest groups.
- other MS NSIs are either in feasibility studies or in partial profiling programs.
- an international study on MNEs aiming at “seeing the whole elephant” was led by Canada in 2005 with the UK, the NL, FR, IT as partners; it tried to check whether the data each country gathered on the same MNEs were consistent; it showed a high degree of inconsistency even for turnover, capitalised expenditure and operating earnings. The only acceptable figure being on employment, the sales variable being correct but difficult to link with the appropriate turnover definition.
2.2 The main conclusion was that the EU had the greatest interest to improve its data consistency and that Eurostat was asked to take initiatives in this field. The MEETS programme offered the right support and the decision was taken to work on improving the accuracy and consistency of
- statistical units for national collection of data
- statistical European aggregates.
This decision was embedded in the ESSnet Profiling (large and complex MNEs), as a part of the EGR program, with the aim of going ahead towards an economic definition of groups and enterprises.
2.3 A definition of profiling is included in the Business Register manual; until now it has suffered no critics, so that it is the base of the ESSnet work:
Profiling is a method:
to analyse the
legal – operational and – accounting
structure of an enterprise group
in order to
establish the statistical units within that group and
their links and
the most efficient structures
for the collection of statistical data
2.4 This definition needs comments:
- the objective is pragmatic and operational, not academic: it consists in defining the most efficient structure for the compilation of statistics, whatever the sources (surveys or administrative data) are; and valid for all European countries;
- the 1st step consists in analysing three types of structures, two of them being classical (legal and operational or managerial), the 3rd being new for many NSIs (the accounting structure). While it is not written, we can guess that the accounting structure is supposed highly related to the existence of reporting procedures.
- the deliverable is clear: establish the statistical units within each group.
2.5 The collaborative work scheme used for profiling is the new European scheme:
-an ESSnet, which is a partnership of NSIs; this one includes 7 partners (CBS–NL, DESTATIS–DE, ISTAT–IT, ONS–UK, SF–FI, OFS–CH , INSEE–FR coordinator), largely representative of the MNEs, as 70% of the large European groups have their headquarters in the partnership countries and developing a common methodology for the whole ESS
- the work programme is structured into 3 steps:
- during 2009-2010 (work done: for more details, please see the following § ):
- a positioning paper on the feasibility of profiling, with stress on the efficiency of the proposals for the collection of statistics and on the acceptability by key users ;
- an input on statistical units as regards to EU regulation eventual updating;
- during 2010-2012: the further elaboration of methodology, tools, guidelines …
-during 2013 and in the following years: a go/no-go decision has to be taken by the concerned Committees of the ESS, then if the go-decision is taken, the full implementation of profiling will take place, inside EU with a related work programme to be built at the beginning of the period (this programme will necessarily include its budgeting and financing means).
- as stressed previously the project is operationally-oriented, so that testing will be part of it throughout its duration: the partners are committed to test on a 4-eyes view (2 partners minimum cooperating on each group), about 40 large Groups (nearly 10% of the intended target of the 500 top MNEs), with the necessary iterations and the production of early results (to avoid the “tunnel risk”: a failure due to duration without result).
3 The profiling options
3.1 An agreement was reached among the partners; it was presented to various audiences without any red light having been identified during this tour. Of course further needs and wishes were expressed and different types of concerns were heard, that the partners think able to be studied in the coming steps of the work.
Therefore they are confident to express the options proposed.
3.2 the 6 following proposals are common to all options.
- necessity of confident relations with profiled MNEs: it seems not possible to profile without reaching an agreement with the profiled MNEs on the Statistical Units that will be surveyed or used in administrative data based statistics and on the principles according to which the data are obtained.
- a Top-Down approach preferred, in the definition of SU:
- because a Bottom-Up approach is impossible when a MNEs includes hundreds or thousands of affiliated legal units,
- because the Bottom-Up approach is quite impossible to keep consistent when the Enterprise Group operates in multiple countries,
- because the accounting “Consolidation” Information System includes a cluster of the legal units of the Group which will be a starting point for analysis of its structure, and because this system is managed in one only point of consolidation.
-because in the EU, the generally accepted accounting principles, the IFRS/IAS system is compulsory for all companies that are listed on the stock and the bond markets and offers the previous characteristics in a consistent way for all he MS.
- and last but not least, because IFRS includes on a compulsory basis information on operational divisions (that we have called “GODS” as a provisional naming), as soon as they are managed on an autonomous basis and have a reporting system, which is close to our definition of the enterprise: the IFRS 8 standard offers this second starting point for profiling and delineating statistical (global) enterprises.
- the introduction of global enterprises (GEs) as the operational divisions or parts of them, are described independently of the country structure of the affiliates,
- Statistical Units and Profiling methodology consistently interlinked so that two kinds of “enterprises” are to be considered:
- the “global enterprises” (GEs): new statistical units usually derived from operational divisions,
- the “truncated enterprises” (TEs) being the national parts of GEs,
leading to the
- need of extra data, as compared to the disclosed “consolidated accounts” e.g:
- intra-group inter-GEs flows (because the sellings and purchases between two GEs have to be described in statistics even if they are suppressed in the global consolidation process).
- intra-GEs inter-countries flows, which have to be included in the national data even if they are also suppressed in the consolidation process.
- National collection of statistics, with the characteristic that national splits of MNEs sum to the whole of it.
3.3 five options for profiling have been explored in the study, with the corresponding benefits,costs and risks; a preferred one has been chosen: the option D «with further consideration to option E».
- Option A is the “no change” option: each country goes on doing as presently
- Option B is the “minor change” option: it works as option A, but tries to apply the present standards in a consistent manner in all Member States; it necessitates important changes to documentation and to the use of it; but does not garranty any kind of improvement in the results.
- Option C is based on the accounting consolidation information system and uses it at the global (highest) level (of consolidation); it delivers country clusters of affiliates; it allows each NSI to define its national enterprises from the country cluster received from the country of the consolidating unit.
This process should be controlled to assess the consistency of all the enterprises. But we did not succeed in explaining what kind of checking were possible without excessive costs.
- Option D supposes that the global groups are profiled by the NSI of the country of the Ultimate Controlling Institutional unit (the UCI, most of the time the country of the headquarters of the Group); then the global group is divided in GEs, then further in national TEs. Results are included in the Euro Groups Register (EGR).
The (truncated) enterprises and, if possible, the list of Legal Units related to them are routed, through the EGR, to all concerned NSIs, for use in their registers and for data collection.
- Option E manages the group cluster, the GEs and the TEs in the same way that the option D; but then data would be collected centrally and apportioned to national levels.
In our proposal, this option is, in the feasibility proposals (§ 13.8), restricted to two cases:
- the highly internationally integrated group (such as aeroplane constructors)
- the groups that would ask to be collected centrally.
4 A consistent system of Statistical Units
4.1 The linkage with the profiling process leads to changes in the definitions of the statistical units and to the following proposals, essential to provide meaningful statistical information:
›Enterprise Groups are
–a set of legal units under common direct or indirect control (mostly a combination of legal units bound by legal and/or financial links)
(they were previously defined as a set of “enterprises”, a circular definition because the EGs were linked to ENT that at their turn were defined as parts of EGs)
›Enterprises produce goods and services, are autonomous for the allocation of their current resources and are either:
–a single legal unit, if it is “independent”
–an enterprise group defined as a set of legal units under common control,
–a part of an enterprise group, with sufficient autonomy.
–SPEs (a special purpose entity or a combination of SPEs)
are to be considered or dealt with as quasi-enterprises
4.2 with a distinction, that constitute the main change, made between:
–Global enterprises, which are the autonomous units, not compulsorily limited within national boundaries
–Truncated (national) Enterprises, which are part of GEs, often not autonomous, but necessary to keep national statistics significant.
4.3 The GEs and the TEs do not have a compulsory type of relationship to legal units:
– (some) legal units may be split into more than one enterprise (that will happen when the enterprise group is organised on a functional basis, e. g when all employees are employed by a single company in a country, or when the R&D is centralised in a specialised affiliate for all the segments of the group etc);
–relations from (GEs & TEs) cluster to (LeU) cluster, a consequence of the previous point;
–we need some further work on how to describe the “rest of world” unit(s): in pragmatic terms it does seem as necessary to describe all the affiliates in the Far East as it is in the EU: but this is a question for the coming future.
4.4 The statistical units constitute a “system”, with highly significant relations and links. This system can be described in the following scheme (Diagram 5 of the report on Statistical units), which has been discussed with the EGR experts since it was elaborated. The reached agreement is that it fits all needs of the EGR and, even more, that it is necessary to make the EGR able to register all types of SU files in the EU context.
Diagram 5 Proposal for generalised multinational Statistical units model
5 An example of the consequences on SBS statistics
5.1 the complete standardisation of Statistical Units at the EU level and data collection with this standard will allow to improve the quality of SBS at EU level.
5.2 The use of “enterprises” with Option D would:
–avoid double counting of variables such as turnover or purchases inside a country, due to intra-group flows (National improvements in SBS and better comparisons possible between countries).
–change industry structure, that would include:
-less support activities, the specialised supporting affiliates being merged with the market-oriented affiliates,
-less services to business, for the same reason , as long as the services to business are sold to other affiliates of the same group,
-more manufacturing activities for the same reason etc
–permit to operate data collection in each concerned NSI, on consistent conceptual and register basis,
–need some more work on classifications: what would be the classification of a truncated enterprise:
-this of the global enterprise
-its own (e.g a distributive activity of the production of the GE or
-dual (permitting to classify at the global enterprise activity or at its own)
5.3the use of IFRS type of data would:
–ensure much better data comparability between countries, because, at least for the largest MNEs, they constitute a set of shared EU standards, more checked and controlled than the present variables definitions
–need to be complemented by intra-group inter-countries flows
5.4Simulations of the quantitative effects of the new rules are planned during 2011
6 Consultation of shareholders and of deciders
6.1 the stakeholders were largely consulted on the previous proposals:
–The BR working group (02/03 & 11/10/2010)
–The FATS WG (01/10/2010)
–The Structural Business Statistics (SBS) working group (16/11/2010)
- as producers of specific data (e.g on concentration), and
- as providers of information for National Accounts
–National accounts, Balance of Payments and Foreign Direct Investments (FDI): invitations to Workshops (as BR, EGR, SBS, FATS) + contacts in MS
6.2the EGR steering group is also used as the Profiling steering group, as profiling is: