The Malmquist productivity index has been based on ratios of radial Fa
rrell efficiency scores. Calculating these relative to piecewise linea
r production frontiers slacks may appear, implying that the full magni
tude of inefficiency is not revealed. Should the index rather be based
on efficiency measures including slacks? Two arguments against are of
fered. Firstly, productivity calculation is different from efficiency
calculation in the sense that the former compares actual data points w
hile the latter expresses hypothetical productivity improvement. Secon
dly, slacks appear due to the ''default rule'' of using ''inner bounds
'' of vertical and/or horizontal segments at the ''edges'' of the tech
nology set.