Ciao am

1
Quality evaluation analysis of the Italian
Business Register on Enterprise Groups Session
Administrative data-Quality and quality assurance
in the use of administrative data Fulvia Cerroni,
Enrica Morganti
2
Quality evaluation analysis of the Italian
Business Register on Enterprise Groups (BR on
Egs) a summary

• main characteristics of the BR on EGs
• quality analysis of the administrative sources
• quality analysis of the process
• quality evaluation of the final output the BR on
  EGs and the quality criteria

3
Main characteristics of the BR on Egs

• Due to the European Regulation N. 2186/93 and N.
  696/93 and in line with the harmonised
  methodology approved and published on
  Recommendation Manual on BR
• Based on administrative sources
• Stock Exchange (STEX)
• Balance Sheet (BAL)
• Chamber of Commerce (CHAMB)
• use of direct and indirect concept of control
• reconstruction of groups through the control
  chains

4
Administrative sources

• Stock Exchange (STEX)
• contains all relevant shareholders of listed
  shared companies
• information structure bottom-up / top down
• Balance Sheets (BAL)
• contains list of consolidated enterprises by
  Italian Group Heads
• information structure top down
• Chamber of Commerce (CHAMB)
• contains all shareholders of non listed shared
  companies
• information structure bottom-up

5
Quality analysis of the administrative sources

• the specific contribute of each source
• preliminary evaluation of matching the sources
• matching rate index and order of matching of
  sources

6
Administrative sources specific contribute
CHAMB
BAL
physical person as ultimate controlling unit
Non resident subsidiaries
Administrative ultimate controlling unit
STEX
7
Administrative sources specific contribute
142.221
6.665
8
Administrative sources matching rate index
Specific matching rate index between the two
sources i,j respect to source i

• Order of matching of the sources

9
Quality analysis of the process

• a view on the process of integration of the
  administrative sources
• Reconstruction of EGs in each source
• Overlapping units analysed by position group
  head/intermediate/terminal units
• Integration of overlapping units
• Integration of units belonging to groups with the
  same group head in both sources
• Integration of units belonging to groups with
  different group head in the sources the case of
  units in different position (figure A)
• quality evaluation of the process of integration
• index of agreement on position
• index of complexity of integration
• results

10
Figure A -The integration of units in different
position
integration
2nd source
1th source
11
Quality evaluation of integration index of
agreement on position
aii
a..
a31
12
Quality evaluation of integration index of
complexity of integration

• The frequency of group head units in one
  source that are in a terminal (a31) or
  intermediate (a32) position in the other source
  can give a measure of how complex the integration
  will be.
• The lower this frequency is, less complex the
  integration would be because there is more
  probability that one source gives ultimate
  controls and the other source will nest in the
  first one.
• Otherwise the higher this frequency is, more
  complex the integration would be because we need
  to go up the control chains in both sources in
  order to identify the ultimate control (figure
  A).
• This confirm there is an order of matching of the
  sources and a priority between them the idea is
  that while STEX has to observe the world of
  listed companies where companies defaulting in
  declarations incur in strong penalty, the other
  sources, BAL and CHAMB, collect information on
  non listed companies where defaulting companies
  are not under control. Therefore knowledge on the
  legal base and empirical evidence give a priority
  order between sources that shows STEX as the most
  reliable of the three input sources, followed by
  BAL and the last one is CHAMB source.

13
The results of administrative source integration

• direct controls
• coverage controls
• profiling activities
• control survey FATS - the survey on foreign
  control enterprises

14
Quality evaluation of the output The quality
criteria relevance

• A statistical business register should be
  relevant to the needs of the users. Therefore
  information collected is relevant if it contains
  relevant units and relevant variables.
• The enterprise group is a statistical unit
  defined in Council Regulation 696/93 on the
  statistical units for the observation and
  analysis of the production system in the
  Community. The increasing diversity of enterprise
  structures in the economy leads to consider the
  enterprise as part of a larger unit, namely the
  enterprise group, where long-term strategies are
  managed in a centralised way.
• Some economic variables are more relevant at the
  group level than at the enterprise level
  particularly operating surplus, research and
  developments taxes, stock exchange
  capitalisation.
• The demand for information on globalisation
  results leads the BR to record information on
  EGs
• EGs can be used to calculate statistical
  indicators for measuring the real size of
  enterprises.
• EGs serve other institutional users e.g.
  Antitrust Authorities, according to dissemination
  policies to calculate statistical indicators for
  market concentration
• EGs provide supplementary information for
  statistical indicators on enterprise demography

15
Quality criteria accuracy

• Accuracy indicates the potentiality of an
  instrument of correctly reflecting reality and is
  considered to be of great importance for the
  quality of statistical business registers. The
  easier way to measure this criteria is through a
  quality survey from this point of view the FATS
  survey, being a direct survey on foreign control
  enterprises, may be considered as a control
  survey on the quality of BR on EGs.
• In the BR on EGs accuracy is achieved if the
  process identifies the population of units
  belonging to groups in a picture as near as
  possible to the real one in terms of
• frequency of enterprises,
• frequency of EGs,
• employed engaged in EGs
• split into the resident and not resident
  populations

16
Measuring accuracy during the main steps of the
process
17
Quality criteria completeness

• A business register is complete if it includes
  all units in the target population.
• To estimate the coverage of the BR on EGs the
  population of shared companies included in the
  Italian statistical BR of active enterprises has
  been taken as target population.
• The estimation of under coverage error in terms
  of size and activity has been measured as the
  weight of estimated shared enterprises on
  theoretical shared enterprises, where the former
  are calculated under the hypothesis that the
  shared companies not included into the CHAMB
  input Archive have the same probability to belong
  to EGs as the included ones.

18
Measuring completeness through the estimation of
coverage error
19
Quality criteria completeness The big EGs as
a good picture of reality
(1) Section J has not been included in the
economic classification Ateco 2002.
20
Quality criteria timeliness and punctuality

• Punctuality measures the speed at which
  statistical business registers are updated to
  reflect real world events. Sometimes it may be in
  conflict with the need for comparability because
  of different timeliness of the administrative
  input data.

21
Quality criteria clarity and accessibility

• Clarity and accessibility concern how easily
  users can access the information held on the
  register.
• This means to give users the possibility to
  access physically to the data and to be able
• to view elementary records directly (for internal
  users only) this is now possible thanks to a the
  existence of a database holding the BR on EGs
• or to access outputs such as summary analyses
  (for external users).
• Accessibility means also to give users the
  instruments to understand information that means
  to release the statistical BR and the appropriate
  documentation necessary to allow users to
  interpret the information correctly (metadata).
  To that end a manual on the BR on EGs is now in
  rough.

22
Quality criteria comparability

• Comparability through space
• Data are collected according to the new version
  of the Regulation 2186/93 which includes and
  specifies which variable must be collected and
  how.
• The Business Register Recommendation Manual in
  Ch.21 Enterprise Groups gives guidelines on
  concepts and definition on EGs, how to implement
  them in operational terms, even though sources
  are different in structure and treatment of
  information follows a common conceptual and
  logical framework.

• Comparability through time

Demography on EGs

• Difficulties are due to the complexity of the
  statistical units.
• From a definition point of view there is yet no
  agreement on how is possible to define that the
  EGs has a continuity during years and if its
  structure is changing in terms of enterprises
  involved.
• We have defined the following basic rule an EGs
  is still the same entity from a continuity point
  of view if and only if it is still having the
  same group head.
• Defects
• Its a restrictive rule,
• problems with non resident group head they have
  no fiscal code so possible only a matching by
  firm name. Work in progress

23
Quality criteria coherence
a. internal coherence rules

• Internal coherence refers to the consistent
  treatment of data within the register, that means
  the consistent application of rules.
• Rule 1 One legal units may have one and just one
  legal unit identified as the first controlling
  unit.
• Rule2 from a definition point of view certain
  units may have a specific position within the
  EGs, i.e. institutions or physical person must
  be the ultimate control unit.

b. external coherence linkage to 1. BR of
active enterprise 2. S13 3. Trade
Register