The October 2022 technical guide
The cover identifies this publication as ACRA Taxonomy 2022 v 1.0, updated in October 2022. It is a software-neutral technical guide for people already familiar with XBRL specifications, particularly data analysts consuming instance documents and vendors developing filing systems. Its architecture and style conventions explain how disclosures become structured data; the taxonomy does not reproduce every accounting requirement. The PDF has 54 physical pages although its printed pagination ends at “54 of 53”.
Disclosure scope and four filing templates
The taxonomy draws on Singapore accounting standards, Companies Act financial statements presented at an AGM or sent to members when a private company dispenses with an AGM, SGX disclosure requirements, Singapore Standards on Auditing, entity/document information, banking requirements and common practice. Its four templates are Full XBRL, Simplified XBRL and financial-statement highlights for banking/finance entities or insurers. Under this guide, Simplified XBRL requires both revenue and total assets not exceeding S$500,000 and no public accountability, together with the signed financial-statements PDF. Other companies use the applicable full or sector-specific route. The banking route covers commercial banks, merchant banks and finance companies; the insurance route covers licensed insurers.
Entry points and package navigation
The filing entry points are sg-sfrs-ent, sg-sim-ent, sg-fsh-ins-ent and sg-fsh-bfc-ent, with date-qualified schema filenames. Select by reporting standards, filing scope and entity type. The combined entry point exists only to view the whole taxonomy and must not generate an instance document. Shared elements, labels, references and formulas feed template-specific presentation, calculation and definition relationships. META-INF/taxonomyPackage.xml identifies publisher, version and entry points; catalog.xml provides OASIS catalogue remappings for offline use.
Modules, prefixes and stable mappings
Core reporting modules separate accounting standards (sg-as), Companies Act (sg-ca), listing rules (sg-lm), auditing (sg-ssa), document/entity information (sg-dei) and highlights (sg-fsh). Separate schemas hold custom types, typed domains, label roles, reference parts and extensible enumerations. The guide’s count table lists 1,835 defined elements: 1,678 accounting-standard elements, 24 company-law elements, 39 entity/document elements, nine listing elements, 56 auditing elements, 25 highlights elements, one enumeration element and three reference parts. Element names and IDs remain stable when terminology changes; labels are updated instead. Interpret labels and linkbases, rather than treating a legacy element name as the current disclosure wording.
Hierarchies, dimensions and extension limits
Simple parent-child hierarchies model most disclosures. Tables such as property, plant and equipment or intangible assets excluding goodwill use dimensions: a table links axes to reportable line items, and axis members identify breakdowns. Explicit members are predefined; typed axes admit values under their domain constraints. The illustrated intangible-assets table separates consolidated/separate information and asset classes, with opening balances, additions, disposals and other movements as line items. There are no tuples. Companies cannot introduce new reportable elements as extensions, preserving comparable filings.
Resource paths and version identifiers
Files may be referenced by relative or absolute paths under the BizFinx taxonomy root. Vendors should reference official files by absolute path and leave them unchanged; permissible linkbase changes are stored separately rather than editing the originals. Namespace URIs identify ownership, module and release, with unique sg-prefixed identifiers. The guide contains inconsistent date examples: its folder instruction specifies 2020-01-01 for the 2022 release, while later examples use 2022-01-01. Use the actual supplied package identifiers rather than mechanically rebuilding URLs from either example.
Restricted values and extensible enumerations
Custom types restrict accounting-standard choices to SFRS(I), SFRS, SFRS for Small Entities, IFRS only or other standards; rounding to actual, thousands, millions or billions; statements to company-level or consolidated; preparation to in-house or outsourced; and financial position to current/non-current or liquidity order. The guide also defines template/version choices, ISO currency codes and a ten-character company UEN format of nine digits plus a letter. Audit-opinion choices distinguish unaudited, clean, emphasis-of-matter, qualified, adverse and disclaimer outcomes. Other fields use Yes/No, Yes-only modified-opinion reasons, or an additional unable-to-conclude/not-applicable outcome where specified. Nature-of-financial-statements extensible enumeration uses the consolidated and separate domain members across all four templates.
Calculation and presentation serve different purposes
Calculation relationships apply weights of +1 or -1 to aggregate or deduct values. They cannot model every relationship, including some combinations of instant and duration facts. Presentation relationships organise disclosures like financial statements: abstract headings group concepts and totals normally follow their components, while a calculation parent appears above its children. Neither a display order nor a label alone supplies a missing calculation rule.
Definition relationships and dimensional defaults
Definition linkbases contain dimensional relationships only, with at most one hypercube per extended link role. Explicit-axis defaults are grouped in one role; hypercubes use scenario contexts and are closed. The consolidated/separate axis is an open dimension available when needed. A separate relationship set, with an empty explicit dimension, supports totals/defaults for facts otherwise reported against typed dimensions. These rules govern valid fact contexts, not merely table appearance.
Label roles, net figures and sign handling
Standard labels identify concepts; terse labels shorten contextual display; total/net labels distinguish aggregation from gross/net calculations; period-start/end labels identify instant balances. Disclosure labels provide useful presentation where ordinary roles do not adequately describe a net or aggregate figure without a defined calculation hierarchy. Negated labels invert visual presentation only. A disposal deducted with weight -1 can be stored as a positive instance value and displayed in brackets or with a minus sign. Changing the label must not change the underlying fact sign or apply the deduction twice.
Authoritative references and their components
Reference roles distinguish disclosure requirements, illustrative examples and common practice. References identify source name and standard number or legislative chapter/section, then paragraph, subparagraph, clause, appendix or example where appropriate. Effective date, expiry date and banking notice are additional reference parts. Source names include SFRS, SFRS for Small Entities, Companies Act, auditing standards, listing rules and Banking Act. Read these references to understand an element’s scope; some entity-information elements and supporting type/domain schemas do not carry references.
Seven categories of formula checks
Formula rules address mandatory facts, conditionally mandatory facts, arithmetic, positive-value constraints, cross-statement consistency, correlated data and dates. An existence assertion checks a mandatory fact such as assets; conditional value assertions require current/non-current components when that statement format is selected. Other examples reconcile opening plus movements with closing balances, compare note detail against statement totals, equate assets with liabilities plus equity, and ensure the period start is no later than the end. Sign checks concern facts that must be positive; expense display conventions do not mean every expense fact must be negative. Rules are divided into common and entry-point-specific files.
Errors, warnings and inactive formula files
Assertions pass when true and fail when false. A genuine error must be corrected before generating the XBRL filing. A possible-error warning can remain an unsatisfied assertion in an accepted filing, for example a query about zero share capital. Missing severity defaults to error; warning assertions link to sev_warning. Formula files prepared for a later implementation phase but not linked from an entry point must not be processed. Follow package references rather than running every formula file found in a directory.
Style precedence and file conventions
Accounting standards, Companies Act and SGX prescribed wording take precedence over the style guide. MUST/MUST NOT are requirements; SHOULD/SHOULD NOT are recommendations. Filenames combine a module prefix, YYYY-MM-DD and a suffix: lab/ref/cal/def/pre for linkbases and elr for role schemas. Namespace prefixes indicate Singapore and their content. The guide expressly allows some prefixes longer than the cited FRTA twelve-character convention where needed for clarity.
Extended link roles and sort codes
Each role has an ID, unique URI, human-readable definition and usedOn attributes. Definitions begin with an eight-digit code in square brackets, ordering filing/disclosure groups before statements and notes. Axis defaults use 99999999. Definitions generally identify a statement, note or axis, with listed exceptions such as filing information and directors’ statements. Roles normally declare use in presentation, calculation and definition linkbases, subject to the axis-default, additional-calculation and cash-flow exceptions. Role IDs describe the relevant statement or note, rather than relying on the sort code alone.
Element names and identifiers
An element ID combines namespace prefix, underscore and name. Names generally use label-based CamelCase, begin with a capital letter, omit special characters and articles A/An/The, and stay within 256 characters. Add nouns to adjectives and descriptors to narratives, dates, counts or percentages. Required suffixes identify Abstract/LineItems, TextBlock, Axis, Table, Member and typed Domain concepts. Do not add Total to element names; use a total-label role instead. Use authoritative references or uncommon abbreviations in names only where needed for meaning, use UK spelling and avoid gender-specific His/Her. Stable-name exceptions remain relevant when labels change.
Readable labels and structural markers
Labels use concise regulatory terminology, initial capitals, ordinary case except proper names/acronyms, UK spelling and prescribed punctuation/hyphenation. Prefer commas to dashes where possible, avoid an Oxford comma in series and normally spell out one to nine except percentages or document references. Keep modifiers in the prescribed order and distinguish current/non-current or gross/net when needed. Narrative and non-monetary labels use meaningful descriptors. Positive/negative alternatives normally use parentheses, while XBRL-only information uses square brackets: abstract, line items, text block, axis, table or member. Standard labels omit Total; total/disclosure roles may use it. Opening/closing labels use “at beginning of period” and “at end of period”.
Element properties and implementation references
Abstract groups, tables, axes and members are abstract and nillable with duration periods; tables use hypercubeItem, axes use dimensionItem and members use domainItemType. The guide assigns boolean and string facts duration periods, and date facts instant periods. New type names end in ItemType and enumerated values use sentence case except proper names/acronyms. Formula filenames and IDs distinguish cross-statement, mandatory, derived-mandatory, correlated, date, arithmetic and sign rules. Reference requirements cover non-abstract elements and dimensional structures, subject to the specific exceptions described earlier. The guide cites IFRS 2019, climate-reporting, FASB and GRI taxonomy documentation for documentation practices; its substantive rules concern the ACRA 2022 taxonomy.
Official source
This article independently explains the substantive contents of the official PDF, including the relevant conditions, procedures and annexes. The linked document remains the authoritative source for its original wording, and later changes should be checked separately.
Read the official PDF ↗