Title: Multilevel path simulation for weak approximation schemes: myth or reality

Abstract: In this paper we discuss the possibility of using multilevel Monte Carlo (MLMC) methods for weak approximation schemes. It turns out that by means of a simple coupling between consecutive time discretisation levels, one can achieve the same complexity gain as under the presence of a strong convergence. We exemplify this general idea in the case of weak Euler and Milstein schemes, and prove that the complexity of the corresponding "weak" MLMC estimates are of order $\varepsilon^{-2}\log ^{2}(\varepsilon)$ and $\varepsilon^{-2},$ respectively. Thus, we propose a new simple approach how to achieve the complexity $\varepsilon^{-2}$ without time-consuming L\'evy area simulation. The numerical performance of the new "weak" MLMC methods is illustrated by several numerical examples.