Over the last 20 years, large claims have been made about the relation between industrial relations and productivity. In the early 1990s, there was wide support for a prediction that displacement of traditional arbitration by enterprise bargaining would unleash a surge of productivity growth. One influential contention was that a 25 per cent increase in productivity was available if wage-setting were decentralised to the enterprise level. More recently, poor productivity performance has been attributed to faults in industrial relations arrangements, especially the power wielded by strong unions. I am sceptical of claims such as these and do not believe that the evidence justifies them.

Productivity describes the relation of production to inputs. A rise in productivity means that more product is obtained from given inputs. In the long run, productivity is the principal determinant of real incomes and hence the standard of living. Hence it is important. There are two standard measures of productivity. Labour productivity is calculated by dividing the value of total output produced in a period by the number of hours of labour expended on it during the same period. Multifactor productivity divides the value of output by a measure of labour hours and capital combined. (In either case, the money values involved are adjusted for inflation.)

If industrial relations arrangements affect productivity, they must do so by altering the amount produced in an hour’s work (labour productivity) or for a given combination of hours worked and capital (multifactor productivity). Alterations to the amounts of inputs will affect production, but not productivity. This is to be borne in mind when it is said (for example) that penalty rates reduce productivity: they may affect production, but productivity is another matter. A recent newspaper column suggested that making it more difficult for people to remain on disability benefits would cause many of them to seek work and thereby raise productivity. It might raise production, but it would not raise productivity. Indeed, it would probably reduce it, since the extra people employed may be less productive than those already in work.

It is conceivable that changing industrial relations arrangements could affect productivity. The employers’ antipathy to arbitration in the early 1990s was based in part on a belief that if employers had more control over workplaces they could get more out of their employees. And strong unions may inhibit management in its efforts to maximise output from given resources. Whether such possibilities are in fact significant is a matter to be determined by evidence rather than arm-chair speculation.

It is usual to measure productivity for the ‘market sector’. This is because it is difficult to measure the output of sectors such as the public service, defence and education, whose products typically do not have market prices. (Efforts are being made to overcome this problem, but they are still experimental.)

The chart below shows the productivity record for Australia between 1964-65 and 2009-10. It covers both labour productivity and multifactor productivity; and in each case it shows both the actual annual observations and the trend lines fitted to them. For labour productivity, the trend line corresponds to an annual growth rate of 2.2 per cent; for multifactor productivity, the growth rate is 1.1 per cent. The faster growth of labour productivity reflects the fact that there is continuous growth in the amount of capital deployed by the average worker.

It is obvious that the trends ‘explain’ much of historical record of productivity. This lends some perspective to the debate about productivity performance. It is difficult to avoid the inference that long-term productivity growth is the fruit of long-term forces, such as technological innovation, rising standards of education, better health and accumulation of capital.

Nevertheless there are variations around the trends. The claim that enterprise bargaining led to a ‘productivity surge’ has to invoke the rise in the productivity curves, relative to trend, in the latter half of the 1990s. The claim that industrial relations have destroyed productivity turns on the experience of the late 2000s.

Neither claim can be conclusively proved or disproven. But there are questions to be posed. In the case of the earlier ‘surge’, why did it peter out? And how can the effects of enterprise bargaining be separated from those of other economic reforms of the 1990s and from those of the increased deployment of IT? For later years, it is hard to relate the claims made to the timing of the productivity variations. As the chart shows, there has been a poor productivity performance since 2004-05. If it were due to the extra ‘clout’ enjoyed by unions under the Rudd-Gillard regime, why did the deterioration set in several years earlier? There is a better fit with the inception of WorkChoices, but the idea of any causal link seems far-fetched.

The Australian Bureau of Statistics has published, on an experimental basis, industry-level analyses of productivity covering the period since 1994-95. Space does not permit me to provide the detailed results. What is clear, however, is that the productivity record differs radically between industries. The overall movements shown in the chart are the weighted averages of quite disparate patterns of experience. In some of the industries industrial relations factors may have played a part, but not in the simple ways suggested by generalisations about managerial discretion and union power.

In the long term, the crucial factors in productivity growth are those noted above: innovation, rising quality of the labour force and growth of the capital stock. In the short term, we need to look behind the whole-economy numbers to find the special factors at work in particular industries. Attempts to locate the causes of good and bad productivity performance in industrial relations are misconceived.

Emeritus Professor Keith Hancock, AO.

Keith Hancock has honorary positions at Adelaide and Flinders Universities where he is studying aspects of regulation and deregulation of the labour market.