XBRL

This article provides a technical overview of the XBRL standard, with emphasis on the architecture of the standard and the potential applications surrounding the business reporting functionality it supports.

Business reporting has been the focus of discussion over the last few years in the wake of increased government oversight. The need for compliance and the need to standardize reporting processes are driving the creation of new applications and standards. XBRL is one such standard that will provide an extensible vocabulary framework within which business reporting can be done. This standard is particularly important due to the broad industry participation in the standardization effort (over 170 members from the financial services, accounting, and technology sectors worldwide) and the increasing evidence of its adoption in practice. For example, in the U.S. the FDIC recommends that all financial institutions post their bank call reports in the XBRL format. The SEC is moving toward requiring that all corporate public financial statements, like 10K filings, be in XBRL format.

XBRL defines a data model for information to be included in business reports, and the basic vocabulary for representing that information in XML. This standard intends to serve a constituency that is spread across a number of industry verticals and across the globe. That brings on an interesting set of challenges. Multiple languages need to be supported for labeling and describing the same business concepts. The same business concepts can be defined and reported in several different ways in different locales. The vocabulary framework also needs to be flexible enough to allow for extensibility with the provision for defining new business concepts and relating them to known standardized concepts. The XBRL specification makes ingenious use of W3C XML Schemas along with XLink to accomplish this task. This article includes a description of how XBRL achieves this flexibility by creating an information model that utilizes XML Schema and XLink to define syntactic constraints for the reporting language, and semantics in terms of relationships between business concepts and associated metadata. The content of the article is based on the latest public draft of the XBRL 2.1 specification.

The XBRL Information Model
The XBRL information model presents us with two components that we can consider: an XBRL instance and its associated taxonomy set. While the XBRL instance contains the concrete facts (e.g., the value of the business concept "sales_per_share") being reported, the taxonomy contains descriptions of the business concepts (e.g., the description of the business concept "sales_per_ share", what its syntax is, how it is calculated, etc.) that are being reported. Figure 1 is an illustration of the information model, and a description of this model follows.

The XBRL instance
An XBRL instance encapsulates business facts through a flexible information model. The most basic components of an XBRL instance are facts, which are pieces of information being reported through the instance. An example of a fact could be "sales in the most recent quarter." A simple fact, which is a piece of information with no structural complexity (for instance, the retail price of a product, a numeric value), is referred to as an item. Related facts that need to be considered together are grouped together in more complex structures as tuples. A tuple can contain items and/or other tuples. Each item can be associated with a business concept, and these business concepts are described and defined in detail through a taxonomy set. The items also reference the context in which they are to be interpreted. The context describes such aspects as the time period to which the fact pertains and the applicable reporting scenario ("actual," "budgeted," "pro-forma," etc.) among other things.

The XBRL taxonomy
An XBRL taxonomy describes a vocabulary by defining the syntax for the terms of the vocabulary using XML Schema, and metadata such as descriptions of the terms and relationships between them using XLink. Each XBRL instance references the schema(s) for its taxonomy, and each schema can reference other schemas. The collection of all the schemas that are required to validate the syntax of an XBRL instance becomes part of that XBRL instance's Discoverable Taxonomy Set (DTS). The XBRL instance can also, optionally, reference XLink linkbases that provide additional information about the business concepts being reported in the instance. XLink is a standard for specifying hyperlinks that link resources. XBRL uses XLink as the mechanism for defining relationships between business concepts and additional metadata (for more information about the XLink standard, visit the W3C XLink page at www.w3.org/XML/Linking). In addition to the XBRL instance, the schemas in the DTS can also contain references to linkbases. Together, all these schemas and linkbases, which describe all the business concepts being reported, form the DTS.

Taxonomy schemas
The taxonomy schemas define the syntactic constraints with which all XBRL instances for that taxonomy should comply. When creating a new taxonomy, you will either start from scratch, or, more likely, you will start from an applicable industry-standard taxonomy and extend it to suit your particular requirements. In either case, this involves importing some schemas and providing your own schema definitions in addition to arrive at the taxonomy schema appropriate for your organization's reporting needs.

The basic XBRL instance schema defines the element types for the XBRL container (the xbrl element), abstract elements for representing items and tuples, and the elements representing contexts. To create a new vertical- or organization-specific taxonomy from scratch, a schema with element definitions for items and tuples related to business concepts in that vertical is created. These element definitions are then placed in substitution groups with the abstract item type and tuple types as their heads. This schema can also include references to appropriate linkbases using W3C XML Schema's annotation facility.

The more likely scenario is that a taxonomy already exists for your industry but you need to perform a little bit of customization, such as adding a few more business concepts to relate them back to the standard concepts. In this case, you will create your taxonomy by importing the industry taxonomy schema and adding new schema components to reflect your custom business concepts. Some industry- or application-specific taxonomies are available at www.xbrl.org/resourcecenter/taxonomies.asp?sid=22.

Taxonomy linkbases
The XBRL specification describes five types of links in the taxonomy linkbases. These links can be organized broadly into three categories: