Retrocausal quantum mechanics (RQM) provides a local causal explanation of Bell correlations. It is widely thought that RQM is consistent with special relativity. In this paper, I point out that this view is not wholly right. It is argued that RQM violates the Lorentz invariance of the temporal relation between cause and effect for certain spacelike separated events in Bell-type experiments.

We review several no-go theorems attributed to Gisin and Hardy, Conway and Kochen purporting the impossibility of Lorentz-invariant deterministic hidden-variable model for explaining quantum nonlocality. Those theorems claim that the only known solution to escape the conclusions is either to accept a preferred reference frame or to abandon the hidden-variable program altogether. Here we present a different alternative based on a foliation dependent framework adapted to deterministic hidden variables. We analyse the impact of such an approach on Bohmian mechanics and (...) show that retrocausation necessarily comes out without time-loop paradox. (shrink)