The Malmquist productivity index is not transitive. This paper argues
that, instead of trying to accommodate this two-period concept to a mu
lti-period setting, it is preferable to look at the measurement of pro
ductivity, efficiency, and technical change in a multi-period setting
from the outset. A new, transitive productivity index is developed wit
h as by-product a measure of the non-neutrality of technical change. T
he various indices are illustrated on panel data of Swedish pharmacies
.