BSDEs with weak reflections and partial hedging of American options. (arXiv:1708.05957v1 [math.OC])

We introduce a new class of \textit{Backward Stochastic Differential Equations with weak reflections} whose solution $(Y,Z)$ satisfies the weak constraint $\textbf{E}[\Psi(\theta,Y_\theta)] \geq m,$ for all stopping time $\theta$ taking values between $0$ and a terminal time $T$, where $\Psi$ is a random non-decreasing map and $m$ a given threshold. We study the wellposedness of such equations and show that the family of minimal time $t$-values $Y_t$ can be aggregated by a right-continuous process. We give a nonlinear Mertens type decomposition for lower reflected $g$-submartingales, which to the best of our knowledge, represents a new result in the literature. Using this decomposition, we obtain a representation of the minimal time $t$-values process. We also show that the minimal supersolution of a such equation can be written as a \textit{stochastic control/optimal stopping game}, which is shown to admit, under appropriate assumptions, a value and saddle points. From a financial point of view, this problem is related to the approximative hedging for American options.