The convergence of moments in the martingale central limit theorem

Abstract

An early extension of the Lindeberg-Feller Theorem was Bernstein's discovery of necessary and sufficient conditions for the convergence of moments in the central limit theorem for sums of independent random variables. In this paper we show that Bernstein's work has a generalisation to martingales. We extend his work in both the independence and the martingale cases by showing that there exists a duality between the behaviour of the moments of the martingale and the behaviour of the sums of squares of the martingale differences. Our proofs are quite unrelated to Bernstein's and are based on Burkholder's inequalities.