First, we consider the problem of hedging in complete binomial models. Using the
discrete-time Föllmer–Schweizer decomposition, we demonstrate the equivalence of
the backward induction and sequential regression approaches. Second, in
incomplete trinomial models, we examine the extension of the sequential
regression approach for approximation of contingent claims. Then, on a finite
probability space, we investigate stability of the discrete-time Föllmer–Schweizer
decomposition with respect to perturbations of the stock price dynamics and, finally,
perform its asymptotic analysis under simultaneous perturbations of the drift
and volatility of the underlying discounted stock price process, where we
prove stability and obtain explicit formulas for the leading-order correction
terms.
Keywords
Föllmer–Schweizer decomposition, simultaneous perturbations
of the drift and volatility, asymptotic analysis, stability