Skip to Content
 Close search

Multifactor Productivity Growth Cycles at the Industry Level

Staff working paper

This paper by Paula Barnes was released on 21 July 2011.

Download the paper

  • Key points
  • Contents
  • Understanding productivity performance and its drivers at the aggregate market sector level requires a closer examination of the underlying productivity performance of individual industry sectors. But interpreting movements in both aggregate and industry productivity measures is not entirely straightforward.
  • Year-to-year changes in measured multifactor productivity (MFP) reflect not only technological progress, but also many other temporary influences.
    • One important such influence is change in the utilisation rate of capital that, because of limited data, is not measured as a change in inputs but instead appears as a change in measured MFP.
  • A common approach when interpreting movements in MFP is to attempt to abstract from these temporary influences through longer-term averaging of measured growth. The Australian Bureau of Statistics (ABS) identifies periods over which to best examine market sector MFP. These are called 'MFP growth cycles' or 'peak-to-peak periods'.
    • The ABS identifies these periods by reference to peak deviations from trend MFP and to general economic conditions at the time.
    • It is potentially misleading to use MFP cycles for the aggregate market sector for analysing industry MFP over time, as the influences affecting deviations from trend vary across industries.
  • Detailed industry studies would be able to provide information to aid in the identification of industry cycles. However, in the absence of such information, the ABS method for identifying MFP growth cycles for the market sector can be made more mechanical and generic so it can be applied to a range of industries. This modified approach includes:
    • a uniform set of rules for selecting which peak deviations are to be used in industry MFP growth cycles - including taking account of the robustness of peak deviations to alternative trend estimates
    • flagging where the selection of specific peak deviations may benefit from further investigation in industry-specific studies.
  • Applying this modified approach at the industry level suggests that there is considerable variation in industry-specific cycles.
    • Less than a quarter of the industry cycles coincide with market sector cycles. For example, industries including Agriculture, forestry and fishing, Mining, and Wholesale trade have no cycles in common with the market sector, while all cycles identified for Manufacturing coincide with those for the market sector.
    • For both Mining and Wholesale trade, a period of negative MFP growth over a market sector cycle is actually a period of positive MFP growth over the closest industry-specific cycle.
  • This approach provides a generic method for the identification of industry cycles, but the paper also notes the scope for further refinement in the identification of cycles where more detailed industry-specific information is available.
  • Preliminaries
    Cover, Copyright, Contents, Preface and Abbreviations
  • Overview - including key points
  • Chapter 1 Introduction and background
    1.1 Background
    1.2 Objectives
    1.3 The rest of the paper
  • Chapter 2 Methodology and data
    2.1 Methodology for identifying MFP cycles
    2.2 Industry MFP data and its properties
  • Chapter 3 Industry MFP growth cycles
    3.1 MFP growth cycles for the market sector
    3.2 Industry MFP growth cycles identified
    3.3 A closer look at industry MFP growth cycles
  • Chapter 4 Implications for analysis of industry MFP growth
    4.1 Comparison of average annual growth in industry MFP over industry cycles and market sector cycles
    4.2 Conclusions
  • Appendix A Additional details about data, methodology and results
  • References

Printed copies

This publication is only available online.

Publications feedback

We value your comments about this publication and encourage you to provide feedback.

Submit publications feedback