Asymptotic lower bounds in estimating jumps (Q395992)

The authors ``study the problem of the efficient estimation of the jumps for stochastic processes. (They) assume that the stochastic jump process \((X_t)_{t\in[0,1]}\) is observed discretely, with a sampling step of size \(1/n\). In the spirit of Hajek's convolution theorem, (they) show some lower bounds for the estimation error of the sequence of the jumps \((\Delta X_{T_k})_k\). As an intermediate result, (they) prove a LAMN property'', that is, local asymptotic mixed normality. The rate of convergence the authors obtain is \(\sqrt{n}\), ``when the marks of the underlying jump component are deterministic.'' Another important assumption is that the process admits only finitely many jumps. The authors ``deduce then a convolution theorem, with an explicit asymptotic minimal variance, in the case where the marks of the jump component are random. To prove that this lower bound is optimal, (they) show that a threshold estimator of the sequence of jumps based on the discrete observations, reaches the minimal variance of the previous convolution theorem.'' Applications include jump estimation in the context of financial high frequency data.