The 'presentist bias' in time-series data: Implications for fisheries science and policy

The bias in catch time series data that occurs when improvements in fisheries catch reporting systems (e.g., consideration of a previously unmonitored fishery, or region) lead to an increase in current catches without the corresponding past catches being corrected retroactively, here called 'presentist bias' is described, and two examples, pertaining to Mozambique and Tanzania are given. This bias has the effect of generating catch time series at the aggregate that appear 'stable' or increasing when in fact catches are declining over time, with potentially serious consequences for the assessment of the status of national fisheries, or in interpreting the global landings data disseminated by the FAO. The presentist bias can be compensated for by retroactive national data corrections as done, e.g., through catch reconstructions.