6.3 Variables

A variable is a user-defined term that links the logical definition in the Data Catalog (DC) to physical definitions in the database referenced by applications. In stress testing, variables act as data inputs that you can modify in a scenario to assess the impact on one or more metrics.

There are different types of variables supported by STSA based on the type of values you want to change. The various types of variables are:

• Categorical Variables - are mostly idiosyncratic variables. These variables may or may not have a pre-determined hierarchy with a fixed set of values. When a variable has a pre-defined set of values, it can be any of the two types:
  • Ordinal Variables- An ordinal variable is a type of categorical variable that can save ordered or ranked values. Examples are rating, asset classification, stage classification and so on.
  • Nominal Variables- A nominal variable is a type of categorical variable that does not have any ordering or ranking among them.
• Numeric Variables-The numeric variables can comprise of general macro-economic variables, micro-economic variables, interest rates, commodity derivatives, other derivatives, idiosyncratic variables and so on. The numeric variables can be broadly categorized as:
  • Term Structure Variables - These variables have a term structure attached to them. They are defined along a term structure with an as of date value for each point in the term structure. You can define the variable as numeric – term structure variable and add a term structure to it. Examples are Interest Rates yield curves, Swap rates, Other Derivatives (Commodity, Currency) and so on.
  • Matrix Variables - These variables are represented in the form of a matrix to showcase movement across a certain ordinal pre-determined hierarchy. Examples are rating transition matrix, credit score transition matrix and so on. While rating and credit score bands are categorical – ordinal variables. A transition matrix represents movement of obligors across the ratings or credit score bands during two time points as a matrix variable. You can define the variable as numeric – matrix variable, and then add a hierarchy to form the metric and also give a value type being a number, percentage and so on.
  • Entity Based Variables - There are numeric variables which are stored in a specific data-entity and are computed or calculated at a relatively lower level of granularity. These variables include Probability of Default (PD), Loss Given Default (LGD), Exposure At Default (EAD) and so on. In case of these variables the level of granularity can be as granular as an account, obligor, or portfolio. The values are stressed and then computed and persisted dynamically during the run execution as a function of the current values (value as of the reference date) and the shock type selected in the scenario.
  • Exchange Rate Variable - These are numeric variable where a currency value is represented in another currency value. For example, US dollar is represented in Indian Rupee (INR). You can give the ‘from currency’ being currency which is measured say USD and the ‘to currency’ being the current which is to represent another currency say INR. So, the variable here is USD – INR with from currency as USD and to currency as INR.
  • Interest Rate Curve - The interest rate curve, also known as the yield curve, is a graphical representation of the relationship between interest rates (or yields) and the time to maturity of debt instruments, typically government bonds. It provides insights into how investors perceive future interest rates, inflation, and economic growth.
  • Computed and Indirect Variables- These variables are variables computed as a function of other variables. They are defined using other variables through an expression or formula.
  • Other Variables:
    • Macro-Economic Variables - These are variables that signal the general trend of the macro-economy. A few examples of these variables are GDP, unemployment, inflation, industrial production and so on.
    • Micro-Economic Variables - These are variables that are industry specific variables such as occupancy rates, air traffic, toll road traffic and so on.

The various types of variable groups are:

• Macro-Economic Variables - These are variables that signal the general trend of the macro-economy. A few examples of these variables are GDP, unemployment, inflation, industrial production and so on.
• Micro-Economic Variables - These are variables that are industry specific variables such as occupancy rates, air traffic, toll road traffic and so on.
• Idiosyncratic Variables - These are variables that can negatively impact individual securities or a very specific group of assets.