If an object is flashed in alignment with a moving object, we perceived the former as lagging behind the latter. There is some agreement that this misalignment (FLE) implies a temporal error: the visual system would ascertain the location of both objects at different times. Consequently, the moving object is perceived leading the flashed one. It is also known that sensory consequences of voluntary actions are perceived earlier than they actually did (Haggard et al. 2002 Nat. Neurosci.). If the FLE implies a temporal error, this would be reduced if the flash were perceived as a sensory consequence of our action. Here we explore this hypothesis. We combined different velocities of the moving object with different spatial offsets between the flash and the object in three different conditions: (1) subjects triggered the flash by pressing a key; (2) the flash was presented after a predictive beep and (3) the flash was unpredictable. We fitted cumulative Gaussians as a function of spatial offset and time. Results can be summarized as follows. First, the time differences predicted the percentage of response-type (p.e. flash lagging behind) much better than the spatial offset. Relative to the unpredictable condition (3), the FLE was reduced about 30% when the flash was internally predicted (1) and 8% when it was predicted externally (2). Furthermore, predictability not only shifted the fitted curves but also decreased the variability. These results are consistent with a spatio-temporal interpretation of the FLE.