Risk Data Taxonomy

Definition

Risk Data Taxonomy is a classification and documentation system for the Risk Data used by an institution.

Usage

A risk data taxonomy might be the organizing principle behind a Risk Data Inventory system.

Context

• It is related but is distinct from the concept of a Risk Taxonomy. A risk taxonomy aims to create a classification system for the risks to which an organization is exposed.
• A taxonomy is less structured than a Risk Data Ontology.

Classification Dimensions

The following are key classification dimensions of a Risk Data Taxonomy

• The Risk Data Types being captured (which in turn links the risk data taxonomy to the overall risk taxonomy). Data types generally reflect the structure of the data:
  • Structured data, which in turn can be in various data structures
    • Timeseries
    • Panel Data
    • Event Data
  • Unstructured data (text, audio, video, images etc.)
• The use of data in risk management / decision making (e.g as risk factors, early warning indicators etc) which reflects the nature of their relationship with risk
• The temporal nature of the data:
  • Static Reference Data
  • Historical data (past risk events)
  • Current state data (latest information available)
  • Projection / Deterministic Forecast (e.g scheduled cashflows at future timepoints)
• The ownership and confidentiality of data
  • Private / Confidential data
  • Public Data (eg Market Data)

The above dimensions are (more or less) domain independent. For any specific risk management area that domain will further induce a classification that is related to the domain. Roughly speaking, every risk type in the Risk Taxonomy will have its own domain related classification dimensions. For example:

Credit Risk Data Subcategories

• Identification Data: e.g., company name, business code, address of registered office, legal form, date of establishment
• Accounting or Financial Data: e.g., company accounting records; Qualitative Data: E.g from review of management structures, interviews etc
• Behavioral Data: Any soft (non-legal) indicators of behaviors / attitudes
• Legal History Data: Any judicial track record of Credit History collected from courts
• Contractual Data (Financial Product data)
• (Asset) Collateral Data and more...
• Reference Data (external identifiers for markets, locations etc.)

Operational Risk Data Subcategories

• Historical Loss Data
• Event Descriptions (Actors, Actions, Assets)
• Key risk indicators and other types of Vulnerability Assessment

Market Risk Data Subcategories

• Reference Data (entities, instruments, markets)
• Prices (bid, offer, or mid-market), opening / closing etc
• Volume, other indicators of activity