You must have to check the stoichiometry of the material. The gap in P-E loop observed may be due to leakage through the material. Such kind of behavior is observed in PZT based ferroelectrics.
Hysteresis test always has preset. There is a delay between preset and measurement, during the delay period decay of state inside the capacitor sample is caused to gap.
There are different answers by different persons above. Someone tells it's due to leakage behavior of the material, someone tells it is due to the time gap between initial state and final state of the switching of ferroelectric domains. Someone tells temporal ferroelectricity phenonenon etc. etc.
But all the answers are summarized in the PDF link sent by Prof. K. Matyjasek. But in this pdf, the hysteresis behavior of the samples are somewhat different type but there is no gap in none of the reported ferroelectric loops.