What can productivity analysis tell us?

The PC has just published and sent me a nice little booklet called the PC Productivity Update. It’s the first of its kind and the new chair Peter Harris tells us in his Foreword that

“Despite the best efforts of statisticians and economists, the measurement and interpretation of productivity remains a challenge. The update seeks to demystify this commonly used, but often misunderstood, concept. Future editions of the update will come out in the March quarter of each year.”

Well I guess the publication demystifies what productivity growth is, though in his tendentious way Ross Gittins does a better job of explaining how it gets hijacked by people who think productivity is whatever might lower their costs. But it’s a very pedestrian trot through the concept of productivity, with little real attempt to take on widespread misuse of the notion. (For reasons further suggested below, I haven’t read it carefully cover to cover so I may be wrong about this – please sing out in comments if I am.)

In any event I’ve always thought that, as commonsensical as it seems that the Productivity Commission focuses as much as it does on analyses of productivity is kind of odd. It’s not odd that it focuses on reports trying to increase productivity as its reports from inquiries do. That’s why the name “Productivity Commission” seemed well chosen when the Industry Commission was renamed. Trouble is, productivity analysis – watching various indicators of productivity as they rise and fall – doesn’t help us improve productivity much. In fact it’s only my natural timidity and equanimity, my weakness of character that stops me from being bolder and claiming that productivity analysis doesn’t help productivity at all. I guess that’s wrong because it is of some interest to know what sectors have had higher productivity growth than others and then to try to work out why. But it’s surprisingly marginal to the endeavour.

Improving productivity, whether it is in the workplace or the nation is about finding things that can be done better, and different pieces of some larger endeavour that could be fitted together better and doing what’s required. Just saying that you want to improve the ratio of inputs to outputs doesn’t tell you anything of use – it’s like saying you want to get rich.  it doesn’t help you work out how to do it.

Here’s the PC on the relevant question (below the fold). I’ll leave them to explain it and Troppodillians to the task of pointing out how the PC’s productivity analysis helps improve productivity. The winner of the competition will be entered in the draw to win a home visit from a panel of eminent Australians (including Paul Kelly and chaired by Tony Jones) to explore the topic “Australian politics: is the party over?” Particular regard will be had to whether we are still the lucky country.

What can productivity analysis tell us?

The Productivity Perspectives Conference featured
examples of analysis of different productivity data.
The headline measure of MFP for the market sector as a
whole gives an indication of productivity performance at
an economywide level, but this broad measure may mask
diversity among different industries. The industry level
productivity statistics published by the ABS (Estimates of
Industry Multifactor Productivity, Cat. no. 5260.0.55.002)
provides detailed industry level data, including capital and
labour inputs and value added and gross output. Analysis
of productivity at the industry level can give economists
particular insight into underlying structural trends in
the economy.

At the industry level, there were two presentations on
productivity in manufacturing. Paula Barnes (Productivity
Commission) examined the proximate causes (that is, the
growth in value added, and in capital and labour inputs)
that were behind the MFP decline in Manufacturing. Harry
Bloch (Curtin University) took a different approach, using
a decomposition of labour productivity in Manufacturing
to capture the effect of capital-embodied technical change.
As well as industry studies, there were also examples
of productivity analysis linking industry-level productivity
performance to broader market sector productivity.

A presentation by Hui Wei from the ABS showed
contributions of each of the 12 market sectors to
aggregate MFP, as well as their contributions to capital
deepening (with and without IT) and to quality adjusted
labour inputs. The indings of the paper suggest that
IT capital deepening has made a sizable contribution to
labour productivity growth between 1994-95 and 2003-04
(Wei and Zhao 2012).

The standard inputs covered by the measure of
multifactor productivity are capital and labour inputs.
Different ways of measuring these inputs can yield
different interpretations. Labour inputs, for instance,
can be measured simply as hours worked, but there is also
a quality adjusted hours worked series in the ABS data,
which takes into account changes in the aggregate quality
of the labour force arising from education attainment and
work experience (ABS 2007). Zhao’s presentation provided
an example of using the quality adjusted hours worked to
estimate the contribution of changes in the labour quality
to growth in gross domestic income