In one embodiment, the method includes selecting a list 1 motion vector of the co-located block in a first reference picture if the co-located block only has the list 1 motion vector. The first reference picture is a type of reference picture permitted to be located temporally before or after the current picture, and the selected list 1 motion vector points to a second reference picture, which is a reference picture of the first reference picture. The method further includes deriving the list 0 and list 1 motion vectors of the bi-predictive block based on the selected list 1 motion vector of the co-located block. This deriving includes deriving a first temporal distance from a difference between the current picture and the first reference picture, deriving a second temporal distance from a difference between the first reference picture and the second reference picture, and scaling the selected list 1 motion vector of the co-located block based on the first and second temporal distances.