What will the election result mean for productivity? Very little

How easily can we really measure productivity increases gained from government policies? AAP

In the first of their debates, Tony Abbott and Kevin Rudd described their policies to boost productivity. Abbott said: “restoring the workplace relations pendulum to the sensible centre, that increases productivity, re-establishing the Australian Building and Construction Commission (ABCC), that increases productivity and that will give our country the stronger economy that we need.” Rudd’s prescription was the National Broadband Network. Neither Abbott nor Rudd said that there were no other proposals to raise productivity; but neither mentioned any.

These were by no means the first claims to have identified the policy sources of better productivity performance. Two that come to mind are the confident assertions of employers, unions and politicians in the early 1990s that enterprise bargaining was the key which would unlock the economy’s productivity potential; and the repeated assertions of the Productivity Commission that microeconomic reform was the magic formula.

My objection to all of these diagnoses is not that they are demonstrably wrong (though the enterprise bargaining prescription comes close), but that they claim too much. The productivity story is complex, and does not lend itself to simple nostrums.

“Explaining” productivity trends

The chart below describes the rise and fluctuations of market-sector productivity over the 47 years from 1964-65 to 2011-12. (Use of the market sector avoids measurement problems that affect the non-market sector.)

The chart distinguishes labour productivity (calculated by dividing production by hours worked) and multi-factor productivity (where capital as well as labour inputs are in the divisor). A logarithmic vertical scale has the effect that a 1% increase occupies the same distance at any productivity level.

Source: ABS catalogue 5204.0. Note that construction of the productivity series has involved linking of earlier and later data.

The straight lines drawn through the two productivity series are trend lines fitted by the standard method. In the case of labour productivity, the trend rate of growth is 2.2% per year; for multifactor productivity, it is 1.1% per year.

The difference exists because much of the growth in labour productivity is due to workers having more capital with which to work. The trends “explain” a large part of the movements in productivity over the whole period — 99% for labour productivity and 95% for multifactor productivity. If we could explain the trends, there would not be much more to say.

Measuring policy interventions

It is logically possible that they are the result of a succession of policy interventions. But that seems unlikely. If policies were the key driver, one would expect rather more jumpy trajectories as successive policies — some more potent than others — were injected and spent themselves.

It is more probable that long-term productivity growth, which dwarfs the short-term deviations, is the product of underlying long-term forces: capital accumulation (physical and human) and cumulative technical progress, the two being intertwined with each other. For these, government is important, but in a long-term sense. Maintenance of law and order and public expenditure on education are conditions of ongoing productivity growth.

The extent to which policies account for the small deviations from trend evident in the chart is a question requiring a more meticulous scrutiny of the record than is possible here. Suffice it to say that policy explanations would jostle with droughts and good seasons and with special factors such as the lag between mining investment and the resultant increase in output.

Remembering history

The claims of Abbott and Rudd ought, in my view, to be assessed with this historical perspective in mind.

In Abbott’s case, there is the problem of interpreting the notion of moving industrial relations to the “sensible centre”. I can only assume that he refers to a range of pro-employer changes to industrial relations law which the Coalition has elsewhere outlined, including more constrained union rights of entry to workplaces, a narrower scope for industrial agreements and easier-to-make Greenfield agreements. Re-establishing the ABCC presumably entails tougher action against misconduct by building unions and their officials.

Will measures of this kind boost productivity? We shall probably never know. All that we can say is that there is so far no evidence that changes in industrial regulation have in the past had any effect on productivity. This is probably because the effects, positive or negative, have been submerged by the more fundamental sources of productivity change to which I have referred. The analogy which comes to mind is throwing a leaf into a river.

Both employers and unions promised that enterprise bargaining would drive a better productivity performance. Two decades later, there is no evidence that it had any such effect. WorkChoices was, of course, a measure to rebalance industrial relations in the employers’ favour.

It was, in fact, accompanied by a temporary slowdown of productivity growth. I do not suggest that it caused the slowdown; but the experience illustrates my contention that changes in industrial relations law have, at best, miniscule effects on productivity.

Technology and productivity

What of Rudd’s claim about the NBN? All else equal, the NBN should enhance the creation and exchange of knowledge, reinforcing the productivity-generating impetus of technical progress. No doubt this is what Rudd meant. The catch lies in “all else equal”. The NBN does not come for free. Rather, it is a major commitment of investable resources which would have alternative uses. Transport infrastructure comes to mind. The positive effects of the NBN may (or may not) be outweighed by the negative effects of forgoing the alternatives. This is an issue that could be pursued in greater depth. For the present, I can only be agnostic.

My tentative conclusion is that the outcome of the election will have little bearing on future productivity performance.