Explanatory Variables

From Open Risk Manual

Definition

Explanatory Variables (also Characteristics, Attributes) is the set of variables (for example ratios, scalar numerical values or boolean indicators) that are used in the context of a statistical Risk Model to explain (and thus potentially also forecast) random behaviour.

Examples

  • Credit Score Factors are typical explanatory variables using in Credit Scoring Models
    • In corporate credit models explanatory variables typically include financial ratios obtained from Balance Sheet / Accounting data such as Profitability, Leverage, Liquidity, Debt Coverage etc.
    • In retail credit models they typically encompass demographic information, wealth indicators etc.

Issues and Challenges

  • There is no scientific means of certifying that the set of explanatory variables is comprehensive (includes all variables with explanatory power). It rests with domain experts to validate this aspect
  • The adjective explanatory puts emphasis on causal relationships, in practice a Characteristic may be used without a clear view on causal influence

See Also