Calibration of stochastic models

Discussion in 'SP2' started by 1495_sc, Jul 22, 2021.

  1. 1495_sc

    1495_sc Ton up Member

    Hello- The core reading for this section is not extensive. I want to understand the practical application of the two approaches for calibration- risk neutral and real word using an example in context of life insurance models (valuation/pricing?). Can someone help? Thank you in advance!
     
  2. Mark Willder

    Mark Willder ActEd Tutor Staff Member

    Yes, this part of the course is deliberately not very extensive. SP2 would not be the place for an extensive discussion of modelling.

    CM2 contains far more detail on these two calibrations and Paper B provides an opportunity to actually use the models.The risk neutral calibration involves using the risk neutral probabilities Q and CM2 gives lots of examples of using this to value derivatives. The real world calibration is the probability measure P in CM2, eg we use the lognormal (geometric Brownian motion) model for shares to be able to work out the probabilities of shares being above/below a certain value.

    So how could the CM2 material be used in practice in SP2? Well as the risk neutral valuation can be used to calculate prices, it can be used for market-consistent valuations in reserving, pricing or embedded values - note we are looking for a single number in the output here, ie the price/value. The real wourld calibration P on the other hand looks at the probabilities of what we think will really happen, so great whenever we want a probability of staying solvent or a probability of a guarantee biting.

    Best wishes

    Mark
     
  3. 1495_sc

    1495_sc Ton up Member

    Thank you, Mark! Its been a while since I revised CM2 material and got a cue from your response. Will build on to it.
     

Share This Page