A Richer Understanding of Australia’s Productivity Performance in the 1990s: Improved estimates based upon firm-level panel data

Australia’s productivity performance is characterized by important differences across continuing firms, frequent entry of new firms, and substantial exit of firms which, for one reason or another, decide to cease production. These basic facts call into question the appropriateness of measuring productivity using an aggregate production function that is based upon a representative firm. This study relaxes the standard assumptions that industries are comprised of a set of homogeneous firms, the set of which are constant over time. Instead, we apply a semi-parametric production to continue production. The model controls for the relationship between productivity shocks and input choices and the inter-relationship between these and the decision to continue production. Using the Business Longitudinal Survey we estimate an improved set of production functions for twenty-five two-digit industries in Australia. We use these results to examine aggregate industry-level productivity performance. We use a new aggregation method in calculating these changes which allows us to separate productivity changes and output composition changes which sheds new light on industry-level productivity performance in Australia.