Measuring performance, development and growth when restricting flexibility

In industries with heterogeneous firms and fast‐changing productivity, it is necessary to consider both firm and time effects if one is interested in technology adoption and efficiency. Since it is not possible to estimate individual firm effects for each time period due to degree‐of‐freedom problems, a trade‐off is necessary with respect to where one is employing a flexible specification and where one is more restrictive. While most studies choose to account for variation, either in cross‐sectional or temporal dimensions, we are in this paper, by clustering firms together in groups, allowing for variation in both cross‐sectional and temporal dimensions. By modifying a shadow cost system, we are able to measure group‐specific temporal patterns of technical and allocative efficiency and technological change without restricting the patterns to any parametric form.