=Paper=
{{Paper
|id=None
|storemode=property
|title=Using Semantic Web Technologies to Facilitate XBRL-based Financial Data Comparability
|pdfUrl=https://ceur-ws.org/Vol-862/FEOSWp2.pdf
|volume=Vol-862
}}
==Using Semantic Web Technologies to Facilitate XBRL-based Financial Data Comparability==
https://ceur-ws.org/Vol-862/FEOSWp2.pdf
Using Semantic Web Technologies to Facilitate
XBRL-based Financial Data Comparability
Héctor Carretié1, Beatriz Torvisco1, Roberto García2
Universidad Rey Juan Carlos
Paseo Artilleros. 28032 Madrid, Spain
{hector.carretie, beatriz.torvisco}@urjc.es
Universitat de Lleida
Jaume II, 69. 25001 Lleida, Spain
roberto.garcia@udl.cat
Abstract. The XML Business Reporting Language (XBRL) is a standard for
business and financial reporting. Many institutions are making available or re-
quiring data in this format, e.g. the US SEC or the Spanish CNMV. However,
XBRL data is loosely interconnected and it is difficult to mix and compare, es-
pecially when it is based on different accounting principles. Our contribution is
based on converting XBRL reports into semantic data and then using Semantic
Web technologies to formalise equivalences among terms from different ac-
counting standards. This approach has been evaluated in a particular scenario
and it is available online.
Keywords. Business, accounting, finance, interoperability, comparability, Se-
mantic Web, ontology.
1 Introduction
There are many attempts to move existing data to the Semantic Web domain, espe-
cially relevant due to the amount of data being mapped are those around the Linked
Data initiative [1]. The main motivation to do so is that usually this data is not offer-
ing its full potential because it is isolated, i.e. not connected to other external pieces of
data that enrich them. It might even be the case that the data is loosely interconnected
internally, because it lacks formal semantics. Most of the time this is due to the fact
that the technological solutions used to publish that data do not make it easy to inter-
connect it internally and to other external data sources.
Business reporting is a domain where the need for a common data format for re-
ports has already been identified. XBRL (eXtensible Business Reporting Language) is
an XML language intended for modelling, exchanging and automatically processing
business and financial information. XBRL is gaining a lot of momentum, especially
thanks to the support of some regulators and government agencies worldwide. It is
especially significant the importance of the XBRL program promoted by the U.S. Se-
curities and Exchange Commission (SEC). Currently, all companies filing to the SEC
are doing so using XBRL following the Government Information Transparency Act,
16
which requires federal agencies to collect their data in a uniform, searchable format
using XBRL thereby simplifying mandatory financial reporting for companies that re-
ceive federal funds.
However, despite the great success in the adoption of XBRL, we have observed
some limitations in its support for cross analysis of financial information in XBRL
tools and applications, as it is detailed in Section 2, that might threaten its usefulness.
These limitations are not just among data based on different accounting principles,
which are represented in XBRL using taxonomies. It even happens when comparing
filings for different companies based on the same taxonomies or filings for the same
company based on different versions of the taxonomies.
We argue that this limitation is inherited from the technologies underlying XBRL,
especially XML. XML takes a document-oriented approach, where each document
presents a tree structure. This makes it difficult for XML-based tools to provide func-
tionalities that blur this separation into documents and that overcome the limitations
of a tree structure when mashing-up data from different sources. Moreover, XBRL
does not provide formal semantics that might help to integrate different taxonomies
using logic reasoners.
In any case, the integration of XBRL data into comparable information is a strong
requirement for the analysis of business and financial information at a global scale.
This might increase the efficiency and effectiveness of the decision-making processes
relying on this kind of information. For instance, bankruptcy prediction and other
tasks related to the assessment of the solvency of a firm, a business sector or set of in-
terrelated companies. Many have already pointed to this issue and propose Semantic
Web technologies as a natural choice for XBRL data integration, cf. Section 2.
Despite these potential benefits, currently, financial and business data is being pro-
duced using XBRL and it seems that more and more XBRL data is going to be avail-
able in the future. XBRL is been promoted by regulators and government agencies
like the US SEC, as it has been shown before, but also other bodies like the European
Union or the Spanish Securities Commission (CNMV) [2].
Consequently, our opinion is that the best short-term approach to enjoy the benefits
of Semantic Web technologies when working with financial data is not to propose and
alternative language based on these technologies, but to apply methods to map exist-
ing XBRL to semantic metadata.
The rest of this paper is organised as follows. The next subsections introduce the
structure of XBRL and Section 2 presents the related work. Then, in Section 3, the
approach for generating semantic data from XBRL is presented. It is based on a trans-
formation from XML data to RDF using the XBRL to RDF mapping, which is de-
scribed in Section 3.1. Then, the second step is to map the XML Schemas that struc-
ture XBRL data to OWL ontologies using the XBRL Schema to OWL mapping
detailed in Section 3.2.
The results of the previous mappings, as detailed in Section 4, are a set of OWL
ontologies for the main XBRL taxonomies used by the US SEC and based on the US
GAAP1. Based on these ontologies, it has been possible to map the XBRL instance
documents sent to the US SEC since 2009 resulting in more than 100M triples availa-
1 Generally Accepted Accounting Principles,
http://en.wikipedia.org/wiki/Generally_Accepted_Accounting_Principles_(United_States)
17
ble from the LOD Cloud as the Semantic XBRL dataset2. Some preliminary experi-
ments have also been done with XBRL data based on the International Financial Re-
porting Standards (IFRS) and the Spanish PGC (Plan General Contable) accounting
regulations.
Section 5 presents the main evaluations done so far. First of all, there are the re-
sults of a basic logical evaluation of the resulting ontologies. Then, we present a
deeper evaluation of the overall approach through an scenario where comparability
between two XBRL reports for the same company but based on different accounting
principles is attained using Semantic Web technologies once they have been mapped
to semantic data. Finally, Section 0 presents the conclusions and the future work.
1.1 XBRL
XBRL is based on two kinds of documents, instance documents and taxonomies. In-
stance documents report business facts and point to a set of taxonomies, which define
the meaning of these facts, e.g. under what accounting principles they hold, what oth-
er facts they related to or what kind of things do they refer to.
1.1.1 Instances
More concretely, a XBRL instance document contains business Facts. An example of
a Fact could be “sales in the last quarter”. If the Fact is simple valued, like “the long
term debt is 350,000” whose value is just a number, it is called Item. If the Fact has a
more complex value, like “for the preferred stock, the preferred stock par value per
share is 0 and the preferred stock shares authorized is 2000”, it is called Tuple.
Items are represented in XBRL as a single XML element with the value as its con-
tent while Tuples are represented by XML elements containing nested Items or Tu-
ples, i.e. subelements.
However, facts are not isolated entities and it is not enough to provide their values,
it is also necessary to contextualize them. Consequently, four more entities are intro-
duced in the XBRL model:
Context: it defines the entity (e.g. company or individual) to which the fact ap-
plies, the period of time the fact is relevant and an optional scenario. The period
of time can have zero length for instance and its value is based on ISO 8601 for
date and time values. Scenarios provide further contextual information about the
facts, such as whether the business values reported are actual, projected or budg-
eted. Contexts are referenced from Facts using the “contextRef” attribute, which
specifies that the given Fact is valid for an entity, period and scenario.
Unit: it defines a unit of measure, such as “USD” or “shares”. They are referenced
from Facts using the “unitRef” attribute, which specifies that the numeric or frac-
tional value of the Fact is based on that unit of measure. Complex units can also
be defined, like “USD per share”. Currency units are based on ISO 4217.
Reference: The kinds of facts under consideration are defined by taxonomies,
which specify their meaning in the context of some accounting principles or pur-
pose, e.g. Facts relevant for banking and savings institutions. These kinds of facts
2 http://thedatahub.org/dataset/semantic-xbrl
18
are then used in instance documents in order to specify actual values for them.
However, they are linked to their definition in the taxonomies, typically through
schema references, in order to be able to retrieve their meaning.
Footnote: it contains some additional support content and it is associated to Fact
using XLink.
Table 1 shows part of an instance document from the EDGAR program that con-
tains a Context element, which defines a company, a time period and the scenario
“unaudited”. Then, there is a Fact that holds in that context. The Fact references the
Context and the value unit, while their content is the fact numeric value.
Table 1. Context and facts examples from an EDGAR filing
…
796343
2008-03-01
2008-05-30
…
54400000
…
1.1.2 Taxonomies
Taxonomies are the other kind of XBRL document. A taxonomy defines a hierarchy
of concepts, basically kinds of Facts, and captures part of their intended meaning. In
XBRL there is a set of base taxonomies that define the core concepts and other ones
that extend them in order to particularize these concepts for concrete accounting prin-
ciples, application domains, etc. Additionally, it is possible to extend existing taxon-
omies and accommodate them to particular needs.
Taxonomies are based on XML Schemas, which provide the taxonomy building
primitives and the extension mechanisms. Moreover, there are also “linkbases”, which
allow establishing links beyond the taxonomy tree structure using XLink.
Schemas define concepts that are instantiated as Items or Tuples, depending on
their complexity, in the instance documents. They are based on XML Schema el-
ements (xsd:element). A concept definition provides the fact name, whether it is a
tuple or an item and its value data type (such as monetary, numeric or textual).
Linkbases define links from concepts in a taxonomy to labels, pieces of content or
other concepts. The XBRL specification defines five different kinds of linkbases.
o Label Linkbase: set of links that provides human readable strings for
concepts, potentially in multiple languages.
o Reference Linkbase: these links associate concepts with citations of
some body of authoritative literature.
19
o Calculation Linkbase: these are links that associate a set of values of
concepts in taxonomies with a mathematical calculation that must be
checked for consistency, for instance that a set of concepts with percent-
age values sum up 100%.
o Definition Linkbase: it provides semantic relations between concepts
like is-a, whole-part, etc.
o Presentation Linkbase: This linkbase associates concepts with other
concepts so that the resulting relations can guide the creation of a user
interface, rendering, or visualisation.
2 Related Work
The U.S Securities and Exchange Commission (SEC) offers some online tools that al-
low interacting with the data available in XBRL form. There is a tool called Interac-
tive Financial Reports that allows viewing and charting companies financial infor-
mation. It also provides some functionality that allows comparing different filings and
different companies, though it is hard to use and prone to even the slightest differ-
ences between the compared filing facts, even when there is just a name change for
facts from filings of the same company.
There is also the Financial Explorer, which presents company financial data
through very informative diagrams but just from one company at a time, and the Ex-
ecutive Compensation tool. The later allows comparing just two facts, Public Market
Capitalization and Revenue, across all filed companies.
Apart from the SEC tools, there are some other XBRL tools, most of them proprie-
tary and with quite high licensing cost. Among them, the Fujitsu XBRL Tools 3 should
be highlighted because they are one of the most popular tool sets and it is available
for XBRL Consortium members and academic users. The tools comprise taxonomy
and instance editors, viewers and validators.
The most powerful tool in this set, though still in beta and with many usability
problems, is the Instance Dashboard. This application can consume multiple instance
documents and, by specifying a base taxonomy, users can perform some comparison
analysis, though limited to facts in a taxonomy that appears in all the filings.
As it can be noted from the previous analysis, the main limitation of XBRL tools is
their limited support for cross analysis of financial information, not just among data
based on different taxonomies, even when comparing filings for different companies
based on the same taxonomies.
This limitation is inherited from the technologies underlying XBRL, especially
from XML. XML takes a document-oriented approach, where each document pre-
sents a tree structure. This makes it difficult for XML-based tools to provide func-
tionalities that blur this separation into documents and that overcome the limitations
of a tree structure when mashing-up data from different sources.
Consequently, Semantic Web tools are being considered by people like Charles
Hoffman, the father of XBRL: “This field [W3C semantic standards] is rich with pos-
3 Fujitsu XBRL Tools, http://www.fujitsu.com/global/services/software/interstage/xbrltools/
20
sibilities and stands as the next logical step in the natural progression of information
technology to seek a higher value proposition” [3].
This interest is materializing, and the combination of XBRL and the Semantic Web
has been receiving some attention in different blogs [4,5], mailing lists and web
groups4. The first attempts to combine both technologies focused on specific for some
parts of XBRL. For instance, there is an ontology about financial information based
on XBRL that is specific for investment funds [6] and, though it is generated using a
generic XBRL taxonomy to OWL ontology algorithm, there is not and equivalent tool
that maps generic XBRL instance data.
Another quite specific tool maps quarterly and semester accounting information
submitted to the Spanish securities commission (CNMV) to RDF [2]. Both approach-
es are based on procedural code specially developed in order to extract specific pat-
terns from the XBRL data. Consequently, they are difficult to scale to the whole
XBRL specification and sensible to minimal changes in it.
More recent attempts have widened and generalised their scope. For instance,
eTEN was an European Community programme providing funds to help make e-
services available throughout the European Union. This programme ended in 2006.
Within this programme there was the WINS project: Web-based Intelligence for
common-interest fiscal Networked Services.
WINS provides a Web-based Business Intelligence (BI) Service to public and pri-
vate Financial Institutions by integrating BI products and knowledge discovery tools
to produce new financial knowledge on companies from information gathered through
interoperable information services. Within the WINS context, Declerk and Krieger [7]
pointed out some limitations encountered in the XBRL schema documents mainly due
to the lack of reasoning support over XML-based data. They proposed the “ontologi-
zation” or process to translate XBRL taxonomies into OWL to overcome these limita-
tions.
The “ontologization” starts from the WINS information extraction (IE) task, which
gathers financial facts from PDF files and converts them into XBRL documents. From
these document, the process continues based on a hand-made translation of XBRL
facts into OWL ontologies that then helps classifying the facts into higher-level con-
cepts like Balance Sheet or Statement of Income. However, the ontologies are not ex-
ploited beyond this point in order to facilitate the comparability of the financial facts
across different accounting standards.
Another example of mapping from XBRL to Semantic Web technologies is
OpenLink XBRL Sponge, which maps generic XBRL instance data to RDF [ 8].
However, in this case, there is not and associated mapping from the taxonomies in-
stance data is based on to ontology languages. Therefore, it is not easy to facilitate the
comparability of the financial facts by working at the conceptual level provided by the
ontologies.
Bao et al. [9] do consider the comparability issue and they point out the tremen-
dous human cognitive effort that must be done when comparing financial data written
in XBRL. Their proposal is to overcome this problem by defining the logic model of
XBRL reports using the Web ontologies language OWL to design ontologies that cap-
4
XBRL Ontology Specification Group,
http://groups.google.com/group/xbrl-ontology-specification-group
21
ture the meaning of the reports beyond just their structure. They transform concepts
into classes and arcroles into properties. However, the possibilities of the logic mod-
els generated are not put into practice in comparability scenarios that involve different
accounting regulations.
Finally, latest approaches start to focus on comparability and attempt to profit from
Semantic Technologies and Linked Data principles to attain it [10]. For instance, the
XBRL European Business Registry (xEBR) is an XBRL Europe project to create a list
of concepts, which are common across the various European business registries. The
concepts encompass basic financial data as well as company profiles. However, this
Project is still limited by the fact that there is no common regulation for Business
Registries in Europe. Therefore, many Registries in Europe have built their own set of
taxonomies.
Our proposal, as detailed in the next sections, focuses on facilitating comparability
at the semantic level, where it is easier to establish the equivalences among financial
facts independently of the particular taxonomies and associated accounting standards
they come from. In order to do that, we propose an approach that, instead of directly
processing XBRL data, takes profit from the fact that it is expressed using XML and
specified using XML Schemas. The instance XML documents are translated into RDF
that models the financial facts and refers to the concepts modelled in ontologies gen-
erated from the schemas. From this point, it is now possible to establish equivalences
that facilitate comparability at the ontology level use inference to benefit from this
knowledge at the instance level.
3 Approach
The proposed approach is based on the transfer of existing XBRL taxonomies and in-
stance data to Semantic Web technologies. This transfer is based on the XML Seman-
tics Reuse methodology [11,12] and the XML Schema to OWL and XML to RDF
tools implemented in the ReDeFer project5.
This methodology combines an XML Schema to web ontology transformation,
XSD2OWL, with a transparent translation from XML to RDF, XML2RDF. The on-
tologies generated by XSD2OWL are used during the XML to RDF step to generate
semantic metadata that takes into account the XML Schema intended meaning.
This approach differs from other attempts to move metadata from the XML domain
to the Semantic Web. Some of them just model the XML tree using the RDF primi-
tives [13]. Others concentrate on modelling the knowledge implicit in XML lan-
guages definitions, i.e. DTDs or the XML Schemas, using web ontology languages
[14,15]. Finally, there are attempts to encode XML semantics integrating RDF into
XML documents [16,17].
However, none of them facilitate an extensive transfer of XML metadata to the
Semantic Web in a general and transparent way. Their main problem is that the XML
Schema implicit semantics are not made explicit when XML metadata instantiating
this schemas is translated. This is so because the RDF data produced from XML in-
5
ReDeFer project, http://rhizomik.net/redefer
22
stance data looses its links to the XML Schemas that structure them and model the re-
lations among different XML entities.
These relations among different XML entities are what carry the XML Schema
implicit semantics. They capture part of the meaning intended by the schema devel-
oper that, though XML Schema does not provide a way to encode semantics, is rec-
orded in the way XML Schema constructs are used. For instance, by modelling that
element “father” is a subtitutionGroup for element “parent”, it is possible to interpret
that “parent” is more general than “father” and that “father” can appear where “par-
ent” appears. More details about the implicit semantics of XML Schema constructs as
compared to OWL ones are provided in Section 3.2.
Therefore, the previous transformations from XML to RDF do not take profit from
the meaning encoded in XML Schemas and produce RDF metadata almost as seman-
tics-blind as the original XML. Or, on the other hand, they capture this semantics but
they use additional ad-hoc semantic constructs that produce less transparent metadata.
3.1 XML2RDF
The XML to RDF transformation follows a structure-mapping approach [13] and tries
to represent the XML metadata structure, i.e. a tree, using RDF. The RDF model is
based on the graph so it is easy to model a tree using it. Moreover, we do not need to
worry about the loss of semantics produced by structure-mapping. We formalised the
underlying semantics into the corresponding ontologies and we will attach them to
RDF metadata using the instantiation relation rdf:type.
The structure-mapping is based on translating XML metadata instances to RDF
that instantiates the corresponding constructs in OWL. The more basic translation is
from xsd:elements and xsd:attributes to rdf:Properties (owl:ObjectProperties for
node to node and owl:DatatypeProperties for node to value relations).
Values are kept during the translation as simple types and RDF blank nodes are in-
troduced in the RDF model in order to serve as the source and destination for proper-
ties. They will remain blank until they are enriched with semantic information.
The resulting RDF graph model contains all that we can obtain from the XML tree.
It is already semantically enriched thanks to the rdf:type relation that connects each
RDF property to the owl:ObjectProperty or owl:DatatypeProperty it instantiates. It
can be enriched further if the blank nodes are related to the owl:Class that defines the
package of properties and associated restrictions they contain, i.e. the corresponding
xsd:complexType. This semantic decoration of the graph is formalised using rdf:type
relations from blank nodes to the corresponding OWL classes.
At this point we have obtained a semantically enabled representation of the input
metadata, a representation that makes the meaning intended by the XML and XML
Schema modelers explicit from a computer point of view. The instantiation relations
can now be used to apply OWL semantics to metadata. Therefore, the semantics de-
rived from further enrichments of the ontologies, e.g. integration links between differ-
ent ontologies or semantic rules, are automatically propagated to instance metadata
thanks to inference.
Focusing on XBRL data, what we get by applying this triplification process of the
corresponding XML data is summarised in Fig. 1. This figure shows the XBRL core
23
concepts as they are modeled in the resulting RDF data. The report is modelled as an
instance of the class “ReportType” and facts are modelled as instances of “FactType”.
In fact, if a direct modelling of the underlying XML tree was performed, facts
should be modelled as RDF Properties because they correspond to XML elements.
However, in order to make the resulting RDF data more usable as it is more intuitive
to view a fact as class instance than as a relation one, we have introduce a modifica-
tion in the basic XML2RDF algorithm as it is detailed in the next subsection.
Then, continuing from the “FactType” instance, there are relations to the actual
value of the financial fact modelled using rdf:value and two properties stating the dec-
imals and unit used for that value. There is also a property linking the fact to its con-
text, which details the involved entity, the time period and the scenario.
Fig. 1. RDF model for the core XBRL concepts generated using XML2RDF and XSD2OWL
(boxes correspond to classes and arrows to properties having them as domain/ranges)
3.2 XBRL Schema to OWL Mapping
The XML Schema to OWL transformation is responsible for capturing the schema
implicit semantics, which is determined by the combination of XML Schema con-
structs. The transformation is based on translating these constructs to the OWL ones
that best capture their intended meaning. These translations are detailed in Table 2.
The XML Schema to OWL transformation is quite transparent and captures a great
part XML Schema semantics. The same names used for XML constructs are used for
OWL ones, although in the new namespace defined for the ontology. XSD and OWL
constructs names are identical; this usually produces uppercase-named OWL proper-
ties because the corresponding element name is uppercase, although this is not the
usual convention in OWL. Therefore, XBRL Schema to OWL produces OWL ontol-
ogies that make explicit the semantics of the corresponding XBRL taxonomies.
The only caveats are the implicit order conveyed by xsd:sequence and the exclusiv-
ity of xsd:choice. For the first problem, owl:intersectionOf does not retain its oper-
ands order, there is no clear solution that retains the great level of transparency that
has been achieved. The use of RDF Lists might impose order but introduces ad-hoc
constructs not present in the original metadata.
24
Table 2. XBRL Schema to OWL translations for the XML Schema constructs
XML Schema OWL Mapping motivation
element[
Facts, though elements, are
@substitutionGroup= owl:Class
mapped to classes
"xbrli:item"]
rdf:Property
Named relation between nodes or
element | attribute owl:DatatypeProperty
nodes and values
owl:ObjectProperty
The corresponding element is
element@substitutionGroup="xb
rdfs:subClassOf mapped to a owl:Class
rli:item"
rdfs:subClassOf xbrli:item
Relation can appear in place of a
element@substitutionGroup rdfs:subPropertyOf
more general one
element@type rdfs:range The relation range kind
complexType|group Relations and contextual re-
owl:Class
|attributeGroup strictions package
Contextualised restriction of a re-
complexType//element owl:Restriction
lation
extension@base | Package concretises the base
rdfs:subClassOf
restriction@base package
@maxOccurs owl:maxCardinality Restrict the number of occur-
@minOccurs owl:minCardinality rences of a relation
sequence owl:intersectionOf Combination of relations in a
choice owl:unionOf context
Moreover, as it has been demonstrated in the Semantic Web community, the ele-
ment ordering does not contribute much from a semantic and knowledge representa-
tion point of view [18] in most cases and when it is a requirement it is more conven-
ient to explicitly represent it using some sort of order attribute or property. For the
second problem, owl:unionOf is an inclusive union, the solution is to use the disjoint-
ness OWL construct, owl:disjointWith, between all union operands in order to make it
exclusive.
4 Results
First of all, we have generated an ontological infrastructure for the XBRL core, cur-
rently XBRL 2.1. It is composed by the ontologies resulting from mapping the XBRL
core XML Schemas using the XBRL Schema to OWL mapping: XBRL Instance,
XBRL Linkbase, XBRL XL and XBRL XLink. These ontologies have been adapted
to accommodate the changes introduced by XBRL to RDF that make the output se-
mantic data more usable, basically by making facts classes and no longer properties.
Apart from the previous schemas, the following schemas have been also mapped in
order to be able to map the XBRL data submitted to the US SEC.
From US GAAP (Generally Accepted Accounting Principles) the schemas, and
corresponding ontologies, are: Primary Terms Elements (USFR-PTE), Primary Terms
Relationships (USFR-PTR), Financial Services Terms Elements (USFR-FSTE), Fi-
nancial Services Terms Relationships (USFR-FSTR) and Investment Management
Terms Relationships (USFR-IME). For specific industries: Banking and Savings In-
25
stitutions (US-GAAP-BASI), Commercial and Industrial (US-GAAP-CI), Insurance
(US-GAAP-INS) and Investment Management (US-GAAP-IM).
There are also some non-GAAP schemas that have been also mapped to OWL on-
tologies: Accountants Report (USFR-AR), Management Discussion and Analysis
(USFR-MDA), Management Report (USFR-MR) and SEC Certifications (USFR-
SECCERT).
The same approach has been followed to map the IFRS taxonomies and the ones
used by the Spanish securities commission (CNMV). Most of the previous ontologies
are available from the BizOntos Business Ontologies web page 6 and the semantic data
for all the processed filings can be queried and browsed from the Semantic XBRL
site7. Currently, more than 25 thousand filings have been processed from the US SEC,
plus some from the CNMV. The combination of all these filings once mapped to RDF
amounts slightly more than 100 million triples. At this step, it is possible to take profit
from semantic web technologies in order to improve the interconnectedness of the da-
taset by means of semantics-enabled data integration.
5 Evaluation
The proposed approach has been evaluated using two input XBRL reports for the
same company but based on different accounting principles, and consequently differ-
ent taxonomies. The input data is from Telefonica S.A., one of the reports was sub-
mitted to the Spanish CNMV and the other to the US SEC8, more specifically the
consolidated Balance Sheet for the years 2009 and 2008.
The motivation is that Telefonica is one of the few Spanish corporations that files
their financial statements to the Spanish securities commission (CNMV) in XBRL
format and also to the American Securities Exchange Commission (US SEC). The
2009 period was the last period available in the CNMV and SEC websites, at the time
of the elaboration of the present evaluation.
The elaboration of the financial statements for the CNMV has been done under the
Spanish GAAP regulations9, i.e. Plan General de Contabilidad, issued in 2007 and
based on IFRS. Meanwhile, financial information filled to the US SEC was elaborated
under the IFRS, following SEC’s provisions for foreign corporations.
Therefore, it could be expected that both XBRL financial reports would be the
same or at least quite similar. However, as the Table 4 shows, there are some differ-
ences mainly due to different levels of disaggregation. The totals for assets or liabili-
ties coincide but not the figures contained under these main sections.
Fig. 2 highlights the accounts where quantity differences are found. For instance,
in the 2009 balance sheet for the SEC (on the left), “Non-current financial assets”
6 BizOntos, http://rhizomik.net/ontologies/bizontos
7 SemanticXBRL, http://rhizomik.net/semanticxbrl
8 Telefonica’s report to the CNMV is available from http://www.cnmv.es/ipps/default.aspx and
the one sent to SEC is available from
http://www.sec.gov/Archives/edgar/data/814052/000095010310000881/dp16939_20f.htm
9 Models recently modified by Ministerial Oder JUS/1698/2011 of June 13, approving the
model for presentation at the Mercantile Registry of the consolidated financial statements
26
amounts 5,988 millions of euros, meanwhile in balance sheet for the CNMV (on the
right) “Inversiones financieras a largo plazo” (long-term financial investments)
amounts 5,499 millions and “Otros activos no Corrientes” (Other non-current assets)
amounts 489 millions. Both accounts sum up 5,988 millions, so the sum of “Inver-
siones financieras a largo plazo” and “Otros activos no Corrientes”, two terms specif-
ic to CNMV taxonomies, is equivalent to the IFRS term “Non-current financial as-
sets”.
Fig. 2. Assets section Telefonica’s Balance Sheet filled to the US SEC
Other equivalences requiring the addition of different account are also highlighted
in Fig. 2 and marked using dark grey. The accounts marked with light grey have di-
rect equivalences between the taxonomies used by the US SEC and the CNMV. For
instance, “Intangible assets” in the US SEC document is equivalent to “Otro inmobi-
lizado intangible”.
The instance document from Telefonica filed to the Spanish CNMV includes terms
specific to the Spanish terminology, defined in the “ipp-gen” namespace in the XBRL
instance documents, but with an equivalent term in IFRS. Other terms reuse the inter-
national standard, and thus are in the “ifrs-gen” namespace, but in some cases they do
not coincide with the terms specified in the IFRS taxonomy. Finally, some elements
are specific to the CNMV, e.g. “ipp-gen:TotalActivoNiif”.
Both, numerical and terminological differences, dramatically decrease the compa-
rability of the two consolidated balance sheets. However, it is possible to establish
equivalences at the conceptual level, and arithmetic operations among them when
there is not a direct equivalence. This is easily achievable thanks to Semantic Web
technologies once the involved taxonomies have been mapped to OWL and the corre-
sponding instance documents to RDF. The next section presents some examples.
27
5.1.1 Mappings between Spanish PGC and IFRS
Table 3 shows some of the semantic mappings generated for the Telefonica scenario
between the ontologies corresponding to the IFRS taxonomies and thus for the
CNMV taxonomies.
Table 3. Mappings between Spanish PGC and IFRS
Spanish CNMV US SEC
Semantic Mappings
(PGC taxonomies) (IFRS taxonomies)
ipp-gen: ifrs: ipp-gen:ActivoNoCorrienteNiif
ActivoNoCorrienteNiif NoncurrentAssets owl:equivalentClass
84.311 € 84.311 € ifrs:NoncurrentAssets
ifrs-gp: ifrs: ifrs-gp:TradeAndOtherReceivablesNetCurrent
TradeAndOtherReceivablesNet TradeAndOtherCurrentReceivables owl:equivalentClass
Current = 10.622€ ifrs:TradeAndOtherCurrentReceivables
ipp-gen:
ClientesVentasPrestaciones CONSTRUCT {
Servicios + [] a ifrs-gp:
ipp-gen: TradeAndOtherReceivablesNetCurrent;
OtrosDeudores xbrli:contextRef ?context;
8.288€ + 2.334€ xbrli:unitRef ?unit;
xbrli:decimals ?decimals;
rdf:value ?value. }
WHERE {
?cvps a ipp-gen:
ClientesVentasPrestacionesServicios;
xbrli:contextRef ?context;
xbrli:unitRef ?unit;
xbrli:decimals ?decimals;
rdf:value ?cvps-value.
?od a ipp-gen:OtrosDeudores;
xbrli:contextRef ?context;
xbrli:unitRef ?unit;
xbrli:decimals ?decimals;
rdf:value ?od-value.
BIND(?cvps-value+?od-value AS ?value) }
The approach is to model the accounts determined to be equivalent, because the are
in the same part of the balance sheet and correspond to the same quantity, as equiva-
lent at the ontology level using the equivalentClass10 OWL construct. When the rela-
tion is more complex than a simple equivalence, for instance when the value for a
term in one vocabulary is the sum of more than one value in other vocabularies, then
the approach is to use a Construct11 SPARQL query that computes the combined val-
ue, for instance the sum, and creates the computed fact.
The complete set of mappings is available from an online demo12, where they are
also put into practice using a Semantic Web repository that includes an inference an
inference and a SPARQL engine that “execute” these mappings. For the demo, just
the CNMV XBRL document is loaded into the repository and the mappings are used
to generate most of the IFRS version of the assets part of the balance sheet using the
semantic mappings.
10 OWL Equivalent Class, http://www.w3.org/TR/owl-ref/#equivalentClass-def
11 SPARQL Construct, http://www.w3.org/TR/sparql11-query/#construct
12 SemanticXBRL Demo, http://rhizomik.net/semanticxbrl-demo/
28
6 Conclusions and Future Work
As it has been shown, it is possible to map the XML data for XBRL filings in order to
generate semantic data that keeps all the original information and structure. This
mapping also includes the involved XML Schemas that structure the XML data. The-
se schemas are mapped to Web ontologies, which make all the semantics implicit in
the original XML Schemas explicit and available when querying semantic data.
Moreover, it is also possible to take profit from Web ontology primitives in order
to semantically integrate different filings following different XML Schemas, i.e.
XBRL taxonomies. Once mapped to ontology concepts and relations, the XBRL con-
texts, facts and other resources defined for different filings can be related as more
specific, more general or equivalent.
This approach has been put into practice in the context of the US SEC’s XBRL
program. It has been possible to apply the previous XML to RDF and XML Schema
to Web ontology mappings to filings sent to the US SEC and some from the Spanish
CNMV. More than 100 million triples have been obtained, which are structured by
the ontologies generated from the corresponding taxonomies.
Moreover, the benefits of the approach have been validated in a real scenario
where it is possible to generate an XBRL report following the IFRS taxonomies start-
ing from one based on the Spanish CNMV taxonomies using semantic mappings es-
tablished at the ontology level.
Future work focuses on, once we establish more semantic mappings at the concep-
tual level that can be reused to map instance documents for different companies, ob-
taining financial statement analysis ratios, taking profit from the semantic data al-
ready available.
For instance, to compute the debt ratio (equivalent to total liabilities / total assets),
and current ratio (equivalent to total liabilities / total assets) by analysing the balance
sheets, or the Return on Sales (ROS, equivalent to net income / sales revenue). From
these ratios and the semantic mapping, we will be able to create a ranking showing
the best-positioned international companies for each ratio mixing the data they submit
to different regulators.
Acknowledgements
The work described in this paper has been partially supported by Spanish Ministry of
Science and Innovation through the Open Platform for Multichannel Content Distri-
bution Management (OMediaDis) research project (TIN2008-06228).
References
[1] Bizer, C., Heath, T., Berners-Lee, T.: Linked Data - The Story So Far. International Jour-
nal on Semantic Web and Information Systems. 5, 1–22 (2009).
29
[2] Núñez, S., de Andrés, J., Gayo, J.E., Ordoñez, P.: A Semantic Based Collaborative System
for the Interoperability of XBRL Accounting Information. Emerging Technologies and In-
formation Systems for the Knowledge Society. pp. 593–599. Springer, Berlin/Heidelberg,
DE (2008).
[3] Hoffman, C.: Financial Reporting Using XBRL: IFRS and US GAAP Edition. Lulu.com
(2006).
[4] Raggett, D. XBRL and RDF. In: Dave Raggett’s Blog, (2008). Available from
http://people.w3.org/~dsr/blog/?p=8
[5] DuCharme, B. Changing my mind about XBRL again. In: Bob DuCharme's weblog,
bobdc.blog, (2008). Available from
http://www.snee.com/bobdc.blog/2008/08/changing_my_mind_about_xbrl_ag.html
[6] Lara, R., Cantador, I., Castells, P.: Semantic Web Technologies For The Financial Do-
main. In: Cardoso, J. and Lytras, M. (eds.) The Semantic Web: Real-World Applications
from Industry. pp. 41–74. Springer, New York, NY, USA (2008).
[7] Declerck, T., Krieger, H.: Translating XBRL into Description Logic: an approach using
Protege, Sesame and OWL. In: Abramowicz, W. and Mayr, H.C. (eds.) Proceedings of the
9th International Conference on Business Information Systems, BIS’06. pp. 455–467. GI,
Bonn, DE (2006).
[8] Erling, O., Mikhailov, I. RDF Support in the Virtuoso DBMS. In: Pellegrini, T., Auer, S.,
Tochtermann, K. and Schaffert, S. (eds.) Networked Knowledge - Networked Media. pp.
7-24. Springer, Berlin/Heidelberg, DE (2008).
[9] Bao, J., Rong, G., Li, X., Ding, L.: Representing Financial Reports on the Semantic Web:
A Faithful Translation from XBRL to OWL. In: Dean, M., Hall, J., Rotolo, A., and Tabet,
S. (eds.) Semantic Web Rules. pp. 144–152. Springer, Berlin/Heidelberg, DE (2010).
[10] O’Riain, S., Curry, E., Harth, A.: XBRL and open data for global financial ecosystems: A
linked data approach. International Journal of Accounting Information Systems. In Press
(2012).
[11] García, R.: Chapter 7: XML Semantics Reuse. In: García, R. A Semantic Web Approach
to Digital Rights Management. VDM Verlag, Saarbrücken, Germany (2010).
[12] García, R., Gil, R.: Linking XBRL Financial Data. In: Wood, D. (ed.) Linking Enterprise
Data. pp. 103–125. Springer, New York, NY, USA (2010).
[13] Klein, M.C.A.: Interpreting XML Documents via an RDF Schema Ontology. Proceedings
of the 13th International Workshop on Database and Expert Systems Applications,
DEXA’02. pp. 889–894. IEEE Computer Society, Washington, DC, USA (2002).
[14] Amann, B., Beeri, C., Fundulaki, I., Scholl, M.: Ontology-Based Integration of XML Web
Resources. Proceedings of the 1st International Semantic Web Conference, ISWC 2002.
pp. 117–131. Berlin/Heidelberg: Springer (2002).
[15] Cruz, I., Xiao, H., Hsu, F.: An Ontology-based Framework for XML Semantic Integration.
Eighth International Database Engineering and Applications Symposium, IDEAS’04. pp.
217–226. IEEE Computer Society, Washington, DC, USA (2004).
[16] Lakshmanan, L., Sadri, F.: Interoperability on XML Data. Proceedings of the 2nd Interna-
tional Semantic Web Conference, ICSW’03. pp. 146–163. Springer, Berlin/Heidelberg,
DE (2003).
[17] Patel-Schneider, P.F., Simeon, J.: The Yin/Yang web: XML syntax and RDF semantics.
Proceedings of the 11th International World Wide Web Conference, WWW’02. pp. 443–
453. ACM Press (2002).
[18] Berners-Lee, T. Why RDF model is different from the XML model. W3C Dessign Issues,
(1998). Available from http://www.w3.org/DesignIssues/RDF-XML.html
30