Reserve Bank of Australia governor Glenn Stevens has ticked off a major item on the 'to do' list for Australian policy-makers: raise the rate of growth in labour productivity.
In his keynote speech at the 2014 ESAMACE conference in Hobart last week, Stevens presented data showing that labour productivity -- the value of output produced per hour of labour time -- increased at an annual rate of 2 per cent in the three years to June 2013.
This is a large jump compared with the rate of just 0.9 per cent per annum in the preceding five years, the period which saw higher productivity growth added to the 'to-do' list. Stevens interpreted the increased rate of growth in labour productivity since 2010 as evidence that "a number of sectors are making serious efforts to control costs and lift productivity".
Labour productivity is up
The argument that the rate of growth in labour productivity in Australia has accelerated in recent years certainly seems robust. For example, looking at slightly different periods than Stevens, in a recent study, I have also shown an increasing rate of growth in labour productivity over the past decade. I calculate that annual labour productivity growth in Australia was 1.15 per cent per annum from 2003 to 2008 and 1.45 per cent per annum from 2008 to 2013.
What I believe requires more scrutiny is the cause of the acceleration in labour productivity growth. To be able to think properly about the causes of growth in labour productivity it is important to understand one big fact: labour productivity differs hugely between industries.
The mining differential
Taking an average over the period from 2003 to 2013, I find that, for example, at the top of the list, an hour of labour in mining produced output worth almost $320, whereas at the bottom of the list, an hour of labour in the accommodation and food services industry produced output worth just under $32 per hour (expressed in 2010-11 prices). Levels of labour productivity in other industries were between these amounts.
The huge differences in labour productivity between industries are mainly driven by differences in the capital intensity of production between industries. Workers in mining have their labour combined with much larger amounts of capital equipment than in accommodation and food services, and hence an hour of their labour produces a larger value of output.
That labour productivity differs so much between industries has major implications for how we should interpret the causes of a rise in the Australia-wide average level of labour productivity.
Ways to interpret the data
One way in which labour productivity can rise is the same as Stevens suggested in his speech. This is if workers within all (or many) industries become more productive. For example, suppose that the value of output produced by an hour of labour in mining increased from $320 to $330 and in accommodation and food services increased from $32 to $33. Then that growth in labour productivity within industries will raise the average level of labour productivity in Australia.
There is, however, a second way in which average labour productivity can increase. Suppose that we take a worker who previously had a job in the accommodation and food services industry and switch their job to the mining industry. That worker used to produce output worth $32 per hour, but will now produce output worth $320 per hour. Hence the average level of productivity in Australia will increase because a worker has been switched from a low-productivity industry to a high-productivity industry.
What I find in my analysis of labour productivity over the past decade in Australia is that it is changes in the industry composition of employment -- the second way I described above -- that have been the main cause of growth. In particular, I find that the increases in the proportion of workers in mining have been most influential.
Between 2002-03 and 2012-13 the proportion of total hours worked in Australia accounted for by the mining industry increased from 1.2 per cent to 3.1 per cent. Because labour productivity is so high in the mining industry, this has the effect of raising the average level of labour productivity in Australia. I estimate that without this shift towards the high productivity mining industry, labour productivity growth in Australia would have been only 0.2 per cent per annum from 2003 to 2008, and 0.1 per cent from 2008 to 2013. So rather than the higher rate of growth in labour productivity reflecting sectors making “serious efforts to contain costs and lift productivity”, I would argue that most of the acceleration in growth simply reflects a faster rate of increase in mining employment.
A different picture
In fact, far from productivity improving within industries, the opposite has happened in some key industries. Between 2003 and 2013 labour productivity in mining fell from $458 per hour to $238 per hour. This is a massive decline and has had the effect of lowering average labour productivity in Australia. I estimate that if labour productivity in mining had remained unchanged from 2003 then the rate of growth in labour productivity in Australia in the past five years would have virtually doubled to 2.8 per cent per annum.
Acceleration in labour productivity growth in recent years therefore should be seen to reflect a shift in employment to the high-productivity mining industry rather than workers within industries becoming more productive. What does this mean for the future of labour productivity growth in Australia?
In the short to medium term, it is probably good news. The mining boom is not over, and while the proportion of workers in the mining industry remains at its current high level, we will continue to get the benefit of having more workers in this high productivity industry. We are also likely to benefit from growth in productivity within the mining industry. The level of labour productivity in mining is currently at an historic low. Falling commodity prices are creating an imperative for cost savings in mining that will drive productivity improvement.
But it is also clear that in the longer-run, we face an adjustment. Once the mining boom slows, and workers shift out of that industry into other industries with lower-productivity jobs, there will be a drag on the average level of labour productivity in Australia. Then we really will need improvements in productivity that come from within industries.