Diapositiva 1

1
(No Transcript)
2
Index

• Good practices and considerations
• Problems
• Future challenges

3
Good practices and considerations
4
Taxonomy driven model

• XBRL taxonomies define
• Business concepts
• Relations between business concepts
• Relations between business concepts and other
  resources
• XBRL is a comprehensive language to express
  business metadata
• Changes in the taxonomy should not impact on the
  system
• Taxonomies as an input

5
Taxonomy driven model example
XBRL Document
HTML Document
6
Taxonomy driven model

• Benefits
• More flexible design
• Reduce errors
• Only structural changes need development

Let the taxonomy be your guide !
7
Taxonomies and data models

• Avoid designing taxonomies based on specific
  views of the data
• Analysts should see the data model behind

8
Taxonomies and data models

• Total incomes by country should be equal to total
  incomes by product
• We need one rule
• TotalIncomes IncomesByProduct
• Calculate total benefits and benefits by country
• We need four new formulas
• TotalBenefits TotalIncomes TotalExpenses
• BenefitsInSpain TotalIncomesSpain
  ExpensesInSpain
• ...
• Add 3 new countries
• We need 9 new concepts
• We need 3 new calculation rules

9
Taxonomies and data models

• With a good design
• Two items
• Incomes
• Expenses
• Two dimensiones
• Country dimension (Spain, Italy, France...)
• Product dimension (TVs, Radios, Fridges...)

countries
items
products
10
Taxonomies and data models

• Total incomes by country should be equal to total
  incomes by product
• This rule is implicit to the model
• Calculate total benefits and benefits by country
• We need to add one formula
• Benefits Incomes Expenses
• Add 3 new countries
• We just need to add 3 new countries

11
Taxonomies based on data models

• Benefits
• Reduce redundancy
• Improve consistency
• Improve maintainability
• Improve flexibility
• The team that develops a taxonomy needs
• Business analysts
• IT analysts

12
Reporting process

• The reporting process is defined by
• Communication channel
• XBRL Taxonomy
• Instance documents
• Additional data

13
Reporting process taxonomy entry points

• Every instance document must refer to
• At least, one schema file
• Any number of linkbase files
• High number of possible combinations !
• Taxonomy entry points constrain the number of
  combinations
• Empty schemas that refer to predefined
  combinations of taxonomy files

14
Reporting process taxonomy entry points

• COREP and FINREP define a multidimensional data
  model
• A set of primary items
• A set of dimensions (and its members)
• The set of possible combinations of primary items
  and dimensions (facts)
• COREP templates define subsets of the whole data
  model
• Some COREP templates overlap

15
Reporting process taxonomy entry points

• Entry points reduce redundancy on instance
  documents
• XBRL representation of data does not depend on
  the template
• Information does not have to be filed twice
• Entry points reduce complexity of the reporting
  process
• Entry points help improving the consistency of
  data
• Entry points join together highly coupled
  templates

16
Reporting process constrains on instance
documents

• Monetary unit
• Precision
• Number of entities and time periods reported in
  one single instance document
• Bank of Spain
• Monetary unit must be Euro
• Precision must be of at least thousands
• One entity and one period per instance document

17
Reporting process additional data

• Taxonomies define business data
• Reporting process need additional data
• Who is the sender of the information?
• When the information was extracted?
• It is important to differentiate process data
  from business data
• Spanish project will use a taxonomy, common to
  every entry point, for this data
• Sender of the information
• Time period reported
• Entity reported
• List of templates submitted
• List of templates not submitted because dont
  apply
• We have avoided to include this information as
  part of the context

18
Reporting process group and individual
information

• Holdings of company groups report information
• Individual as a single company
• Consolidated as a group of companies
• Two alternatives
• Including one additional dimension individual /
  consolidated
• Considering the group as a different entity
• Single company has code 0456
• The group has code 0-0456

19
General project considerations

• XBRL as the only format available to file data?
• XBRL scares people
• Several choices
• Allow more than one format
• Easiest for financial entities
• Two formats must be maintained

20
General project considerations

• XBRL as the only format available to file data?
• XBRL scares people
• Several choices
• Allow more than one format
• Only XBRL
• Key establish a fluent relationship with banks
  software providers

National Bank
xbrl
xbrl
xbrl
Financial entity
Financial entity
Financial entity
21
General project considerations

• XBRL as the only format available to file data?
• XBRL scares people
• Several choices
• Allow more than one format
• Only XBRL
• XBRL is the only format allowed, but providing a
  translation tool
• Very similar to German solution
• Entities to consider providing XBRL directly
• Translation tool must be maintained

National Bank
xbrl
xbrl
xbrl
Financial entity
Financial entity
Financial entity
txt
22
General project considerations

• Error management
• Expect information with errors
• Design carefully how errors and exceptions are
  handled
• Dont impose constrains that are not needed
• Notify errors with clear user messages and
  suggested corrections
• Budget support resources
• Use log files to store errors and warnings

23
General project considerations

• Providing helper services
• Instance validation tool
• Instance visualizing tool
• Taxonomy dictionary
• Taxonomy browser
• Or relying on third parties to provide this
  services

24
General project considerations

• Software platform
• Validation
• Visualization
• Instance creation (e-forms)
• ETL
• XBRL market tools vs. in-house development
• Spanish COREP
• Commercial XBRL engine for validation
• Visualization and ETL based on XSLT, but using
  the API of the XBRL engine

25
General project considerations

• Project deployment
• Communication plan
• Appoint meetings with financial entities
• Appoint meeting with software providers
• Real data for testing
• System implementation
• Taxonomy implementation
• Adequacy of business rules
• Performance
• Support
• Budget support resources

26
Problems
27
Closed and open reporting models

• The closed model constrains the data to be filed
• What MUST be filed
• What CAN be filed
• Any other information cannot be filed
• Regulators following a closed reporting model
  dont allow extra information to be sent
• The open model just establishes a set of
  guidelines
• Entities can extend the information

28
Closed and open reporting models

• XBRL does not support some basic constrains
  required by a closed reporting model
• Data that must be filed
• An empty instance document is valid
• It will be possible to express these constrains
  with the future formula linkbase
• Possible adequate ?

29
Calculation linkbase not enough

• Calculation linkbase allows definitions of
  summation items networks
• Other basic arithmetic operations are not allowed
• With formula linkbase will be possible
• Summation across dimension members is not
  possible
• With formula linkbase will be possible
• Possible adequate ?
• Calculation linkbases are extensible
• Formulas linkbases will be
• But individual formulas wont

30
Calculation linkbase not enough

• Summation checks are run when an instance
  document contains at least one value for each
  part of the equation
• A B C D

31
Modularization

• Modularization should make easier reusing
  available taxonomies
• Instead of importing the whole taxonomy, you just
  import a part
• It is the author of the extended taxonomy the one
  that decides the criteria to split the taxonomy
• COREP modular design
• One taxonomy per dimension (14)
• One primary items taxonomy per business template
  (18)
• Some common primary items taxonomies (3)
• One template taxonomy per business template (18)

32
COREP dependencies graph of t-mi template
33
COREP dependencies graph of t-mi template Spanish
extension
34
Modularization

• Is modularization something users should be
  really care?
• Should this modularization issue be solved by
  XBRL technology instead of taxonomy designers?
• Have a look at the presentation Large
  Taxonomies, Small Footprint by the Fraunhofer
  Institute at the Madrid XBRL International
  Congress

35
Challenges
36
Evolution of the SIIF project
Bank of Spain
Financial entities
FIN (Mainframe)
SIIF (XBRL)
37
Evolution of the SIIF project

• Improve the quality of data
• Validation at the origin
• Accurate definition of data
• Flexibility (taxonomy driven model)
• Better data model better analytical capabilities

38
Inter-instance validation

• Validations between different time periods
• Validations between different statements
• Validations with data from other systems
• Current calculation and formula linkbases work on
  instance document level
• Inter-instance validation is not considered

39
Inter-instance validation

• One possible solution
• Produce auxiliary instance documents that put
  together information from different time periods,
  statements, ...
• Run validation on these instance documents
• These instance documents would be defined by
  internal taxonomies
• Performance?
• Maintainability?

Financial entities
XBRL
Instance builder
40
Inter-instance validation

• Validation engine with storage integrated
  solutions
• Taxonomy designers defines rules
• XBRL engine knows where to get proper instance
  documents
• State of the art of technology?

41
XBRL Storage and operation

• Which is the best solution to store XBRL data?
• Relational databases?
• XML databases?
• Mixed solutions?

42
Versioning

• How to minimize the impact of new versions on our
  systems?
• How to deal with new versions of taxonomies
  imported by our taxonomies?
• How to compare data defined by different versions
  of the same taxonomy?

43
Link role registry

• XBRL taxonomies define
• Business concepts
• Relations between concepts
• There is a set of standard relations
• Calculation linkbase
• Reference linkbase
• Presentation linkbase ...
• But XBRL allows user-defined arc-roles
  user-defined relations
• XBRL Internationals Link Role Registry to
  include user-defined arc-roles
• Efficient way to extend XBRL without touching
  core specification
• Should we promote to the link role registry
  arc-roles that cover our common needs?

44
(No Transcript)
45
Design stage business rules

• Current XBRL specification has limited validation
  capabilities
• XBRL Formula will not be available in short time
• Possible solutions
• Define our own implementation of formulas
• XBRL allows user defined arc roles and attributes
• Flexible, but expensive solution
• Use a propietary formula implementation
• State of the art of formula and dimensions ?
• Develop specific code to solve your current needs

46
Design stage taxonomy driven model

• XBRL taxonomies define
• Business concepts
• Relations between business concepts
• Relations between business concepts and other
  resources
• XBRL is a comprehensive language to express
  metadata
• Changes in the taxonomy should not impact on the
  system
• Taxonomies as an input
• More flexible
• Reduce errors
• Only structural changes need development

Let the taxonomy be your guide !
47
Design stage cache issues

• High number of documents in short time
• XBRL validation can be a heavy process
• But, loading XBRL taxonomies is the more
  expensive one
• Cache mechanisms
• Entry points simplify the problem

Number of documents received
48
Reporting process

• COREP defines a closed reporting model
• COREP templates reject facts not expected
• But XBRL (using standard arc roles) cannot check
  data that MUST be sent.

49
General project considerations

• XBRL as the only format available to file data?
• XBRL scares people
• Several choices
• Allow more than one format
• Easiest for financial entities
• Two formats must be maintained

50
General project considerations

• XBRL as the only format available to file data?
• XBRL scares people
• Several choices
• Allow more than one format
• Only XBRL
• Key establish a fluent relationship with banks
  software providers

National Bank
xbrl
xbrl
xbrl
Financial entity
Financial entity
Financial entity
51
General project considerations

• XBRL as the only format available to file data?
• XBRL scares people
• Several choices
• Allow more than one format
• Only XBRL
• XBRL is the only format allowed, but providing a
  translation tool
• Very similar to German solution
• Entities to consider providing XBRL directly
• Translation tool must be maintained

National Bank
xbrl
xbrl
xbrl
Financial entity
Financial entity
Financial entity
txt
52
Design stage visualization service example
XBRL Services
XBRL engine API
XBRL
HTML
Front-end
XBRL
HTML
Financial entity
53
Design stage visualization service example
XBRL engine API
Front-end
XSLT Processor
XBRL
HTML
Financial entity
54
Design stage visualization service example
Front-end
XSLT
Javascript
XSLT
Javascript
Financial Entity
XSLT Processor
55
Reporting process extending the European taxonomy