In this paper I discuss and evaluate different arguments for the view that the laws of nature are metaphysically necessary. I conclude that essentialist arguments from the nature of natural kinds fail to establish that essences are ontologically more basic than laws, and fail to offer an a priori argument for the necessity of all causal laws. Similar considerations carry across to the argument from the dispositionalist view of properties, which may end up placing unreasonable constraints on property identity across possible worlds. None of my arguments preclude the possibility that the laws may turn out to be metaphysically necessary after all, but I argue that this can only be established by a posteriori scientific investigation. I therefore argue for what may seem to be a surprising conclusion: that a fundamental metaphysical question – the modal status of laws of nature – depends on empirical facts rather than purely on a priori reasoning.