Chapter 9: Stochastic calculus and Ito processes Flashcards
1
Q
Alternative name for Geometric Brownian Motion
A
Continuous Time Lognormal Model
2
Q
Why might Geometric Brownian Motion (a.k.a Continuous Time Lognormal Model) not be appropriate for modelling share prices
A
- σ may not be constant over time (estimates of volatility from past data depend on the time period chosen)
- μ may not be constant over time (e.x. influenced by interest rates)
- Mean-reversion, as seen in some markets, are not consistent with the assumption of independent increments.
- Momentum effects “ “ “ “ “.
- Distribution of observed log returns has a smaller peak than implied by the normal distribution.
- The distribution of observed log returns has fatter tails than implied by the normal distribution.