How does the brain match corresponding inputs across sensory modalities? Neurophysiological and behavioural studies have suggested that multisensory integration is facilitated by spatial and temporal overlap between inputs (the so called spatial and temporal rules). These rules are sensible, because both cues suggest strongly that the multimodal signals arise from a single cause. A similar inference is reasonable when we ourselves produce an action which causes reafferent feedback in two or more sensory modalities, as when we bang a drum. This observation suggests the possibility of an efferent rule for multisensory integration. We sought experimental evidence for the existence of such a rule. Experiment 1 made use of temporal order adaptation, the tendency for observers to adapt to a multimodal temporal asynchrony following repeated presentations. We predicted that adaptation should be greater for adapting stimuli that could be reliably paired based on efferent information. 18 participants adapted to three asynchronies (0, -200, +200 ms) of a spatially overlapping sound-light pair, and were tested in occasional temporal judgement trials to assess how their points of subjective simultaneity changed. Causality was manipulated in three conditions (causal: button press produces stimulus pair after a short, constant delay; non-causal: button presses are made, but dissociated from the stimulus pair using a longer variable delay; control: No button presses, just stimulus pairs). Simultaneity judgements revealed reliable adaptation, with a similar but non-significant trend in order judgments, but adaptation did not differ across the three experimental conditions, providing no evidence for an efferent rule. Two further experiments assessed multisensory binding more directly, by measuring the variability of judgements about temporal order and simultaneity in response to sound-light pairs coming from the same or different spatial locations. In line with previous results, judgements were more uncertain when stimuli came from the same spatial location, consistent with the stimuli having been combined into a single event. However, we were unable to demonstrate a similar effect when comparing pairs occurring after a key press with those arising at a predictable time but without a preceding action, at least when possible confounds were considered. Hence our search for evidence of an efferent rule in multisensory binding has not yet yielded positive results.