1: \begin{abstract}

2: 

3: We consider backward stochastic differential equations (BSDEs) with mean-field and McKean--Vlasov interactions in their generators in a general setting, where the drivers are square--integrable martingales, with a focus on the independent increments case, and the filtrations are (possibly) stochastically discontinuous.

4: In other words, we consider discrete- and continuous-time systems of mean-field BSDEs and McKean--Vlasov BSDEs in a unified setting.

5: We provide existence and uniqueness results for these BSDEs using new \textit{a priori} estimates that utilize the stochastic exponential. 

6: Then, we provide propagation of chaos results for systems of particles that satisfy BSDEs, \textit{i.e.} we show that the asymptotic behaviour of the solutions of mean-field systems of BSDEs, as the multitude of the systems grows to infinity, converges to I.I.D solutions of McKean--Vlasov BSDEs.

7: We introduce a new technique for showing the backward propagation of chaos, that makes repeated use of the \textit{a priori} estimates, inequalities for the Wasserstein distance and the ``conservation of solutions'' under different filtrations, and does not demand the solutions of the mean field systems to be exchangeable or symmetric.

8: Finally, we deduce convergence rates for the propagation of chaos, under advanced integrability conditions on the solutions of the BSDEs.

9: 

10: \end{abstract}

11: