Real wages, inflation and labour productivity in Australia

Abstract

This article presents an analysis of real wages, inflation and labour productivity interrelationships using cointegration, Granger causality and, most importantly, structural change tests. Applications of tests to Australian data over the 1965 to 2007 period corroborate the presence of a structural break in 1985 and show that a 1% increase in manufacturing sector real wages led to an increase in manufacturing sector productivity of between 0.5% and 0.8%. Comparable estimates for the effect of inflation on manufacturing sector productivity have limited statistical significance. Granger causality test results suggest that real wages and inflation both Granger cause productivity in the long run.

Keywords:

• labour productivity
• real wages
• inflation
• cointegration
• Granger causality

JEL Classification::

• C50
• E23

Acknowledgements

The authors are grateful to Professors B. Bhaskara Rao, Dimitris Christopoulos, Erinc Yeldan, Jack Strauss, Peter Howells and Toni Mora for comments on earlier drafts. All errors are the authors’ responsibility.

Notes

¹ In addition to those discussed elsewhere in this article the list includes: Geary (1976), Denny and May (1977), Gordon (1984, 1988), Buck and Fitzroy (1988), Crafts (1992), De Gregorio (1992), Sbordone and Kuttner (1994), Smyth (1995), Cameron et al. (1996), Feijo (1997), Palokangas (1997), Hondroyiannis and Papapetrou (1998), Freeman and Yerger (2000), Blanchard and Katz (1999), Fountas et al. (1999), Ghali (1999), Tsionas (2003b), Dritsakis (2004), L’horty and Rault (2004), Hsu (2005), Mahadevan and Asafu-Adjaye (2006) and Bildirici and Alp (2008).

² Our productivity measure is output per hour per worker which is calculated as total industrial output (in monetary terms) divided by the total industrial employment. Nominal wages are deflated by the consumer price index to provide a measure of workers’ real purchasing power.

³ For more details on these tests, see Rao and Kumar (2007, 2009) and Kumar (2009).

⁴ See Gregory and Hansen (1996a) for the tabulated critical values of cointegration tests with unknown breaks using the EG method.

⁵ Due to short sample biases, we ignored the sub-sample estimations. However, in any other situation where quarterly or monthly data is used, it would be prudent to estimate cointegrating equations using sub-sample periods considering the break dates from the structural break tests.

⁶Further investigation is necessary to identify whether this is due to: (1) higher real wages increasing the opportunity cost of job loss that can stimulate greater work effort to avoid redundancy, (2) higher real wages putting upward pressure on labour costs and causing firms to substitute labour for capital, thereby increasing in the marginal productivity of labour, (3) greater capital stocks increasing the demand for (typically high skilled) labour, thereby increasing the real wage, and stimulating productivity or (4) domestic pressures on real wages stimulating a movement towards the adoption of capital.

⁷It could also be because, it should be inflation expectations rather than actual inflation which drive investment decisions and therefore future productivity.