Randomised scenario generation

Delivery of a procedure and an application module calibrated by machine learning, making it possible to synthesise an infinite number of macro-economic, commercial and financial scenarios: joint generation of plausible trajectories of statistically linked indicators.

Randomised generation

• Exploration of an infinite - but probabilised - field of possibilities, made up of credible trajectories, more or less stressed,

• Calculating an estimator of the probability distribution of any profitability or risk indicator in this field,

• Conducting stress tests on this field.

• However, unlike the projections carried out as part of the financial steering process, this is a joint projection of the trajectory of both macro-economic indicators and commercial and financial performance indicators.

Macro-economic scenarios

These could include, for example

• macro-economic parameters (GDP, inflation, unemployment, etc.)

• exchange rates and market yield curves,

• customer rates (loans and deposits)

• production volumes (loans and deposits),

• balance sheet outstandings (loans, deposits, capital, securities, balance sheet total, etc.)

Example: interest rates and sight deposits in France