The statistical struggle to get productivity right
Productivity has emerged as probably the totemic issue of the Australian economy in recent years. Many in business, academia, politics and the media believe that the country has fallen behind competitors, that our workers are high cost and unproductive and that our industrial relations laws prevent business from fixing that. We need to fix productivity, they say, or our recent rise in living standards will come to a halt.
Some of us have taken issue with some or all of these arguments, arguing that productivity has been better than believed, or that our performance in the 1990s and 2000s was poorer than we thought, or that the reasons for poor productivity growth are one-offs, such as the huge mining investment boom, and that it will recover as mining production ramps up as a consequence of new investment. Moreover, one area of productivity has been improving. The national accounts for the March quarter showed that productivity (GDP per hour worked) rose 2.1% in the 12 months to March, and gross value added (another measure) was up a solid 2.4%. Both were much stronger than the figures for the 2013 calendar year of 1.7% and 1.8% respectively. A policymaker who has previously argued for the importance of productivity growth, Reserve Bank governor Glenn Stevens, acknowledged the recent rebound in his big speech in Hobart at the start of this month.
Many in business and the commentariat are loath to acknowledge improved productivity and have also failed to admit the recent slide in real wages is also delivering big benefits for businesses — Stevens was optimistic that even if the value of the dollar fell, wage rises would not add to inflationary pressures — another argument that many in business and commentariat have failed to acknowledge.
The Productivity Commission has also noted that labour productivity in some public sector occupations, such as health and education, is difficult to measure — which is particularly a problem given the health and social care sector is the biggest employer of Australians and growing rapidly as we age.
Productivity of course isn’t merely about labour productivity — even as labour productivity has recovered, Australia has experienced a slide in multi-factor productivity (MFP) in recent years (an experience shared with practically every major developed economy, especially the United States and United Kingdom). MFP refers to the addition of capital to labour to produce a total picture of the performance of businesses across the economy and its various sectors. In mining, for example, “mining capital services growth” — the rise in the level of capital investment — hasn’t yet led to an appropriately large increase in output, meaning MFP has fallen significantly.
But measuring MFP is far more difficult than measuring labour productivity, as the PC has noted on numerous occasions. Economists are always looking for ways to “bring productivity closer to its conceptual definition”.
Later this year, the Australian Bureau of Statistics launches a trial in the December quarter’s national accounts (which will be released in early March 2015) that will attempt to measure the MFP of the mining industry. Accurately measuring productivity in that sector (which now accounts for around 9% to 10% of the economy) is particularly problematic when the sector has been undergoing such dramatic expansion as we’ve seen in the last decade. The ABS says “2012-13 was the ninth consecutive year of negative or very weak MFP growth in the market sector. This pattern over the last decade is atypical in the longer-term history of Australia’s productivity performance.”
The ABS plans to include estimates of mineral and energy resources in its estimates of mining industry productivity - but continue to offer estimates exclusive of it as well, to enable consistent comparison. It turns out that when those estimates are included, “mining capital services growth is significantly reduced when mineral and energy resources is included, thus reducing the decline in measured MFP for the mining industry”. In fact, it reduces it quite significantly, as an accompanying graph shows:
Page 1 of 2 | Next page