QM016 Hands-on Lab
EWMA-t Scenario Simulation
Objective
Estimate a current EWMA covariance state from synthetic returns, simulate multivariate Student-\(t\) shocks with the same target covariance, and compare tail frequency with a Gaussian baseline.
Run
python python/qm016_ewma_t_simulation.pyCheck
The Student-\(t\) sample covariance should be close to the EWMA target when the scale conversion is correct. Its large standardized shocks should occur more often than under the Gaussian baseline.
Experiment
Change df from 5 to 8 and 20. Then change lam from 0.94 to 0.97. Separate the effect of tail thickness from the effect of the current covariance state.
The data are synthetic and are not evidence for any empirical parameter choice.