Evatt Q&A on productivity

What does ‘productivity’ mean?

‘Productivity’ has two distinct but easily confused meanings, which is the source of a good deal of misunderstanding in public debate.

In everyday usage, ‘productivity’ refers to being more or less productive, usually by generating higher or lower levels of output. The second meaning is technical and specific to the discipline of economics. In this strictly limited usage, ‘productivity’ doesn’t merely refer to producing more or less, but to producing more or less as a ratio of what is consumed in the act of so doing.

Can you illustrate the difference with a simple example?

‘Productivity’ in the everyday sense has a loose or non-prescriptive meaning that can be applied in countless ways. If someone says ‘my productivity was high today’, they usually expect to be understood as saying that they ‘produced a lot’. If an economist says ‘my productivity was high today’, they could be saying that they actually produced less but used even less energy and materials to do so.

Similarly, if someone says that their lemon tree has high productivity, they probably just mean that their tree produces a lot of fruit. But if an economist found that the tree was consuming high levels of fertilizer, its productivity might be low.

In everyday use, ‘productivity’ just means being productive. In the discipline of economics, ‘productivity’ is not a measure of production but the ratio of ‘output’ over ‘input’.

Coleridge: ‘not, indeed, as the product, but as the producing power — the productivity’.

To draw the distinction very finely, the romantic English poet, Samuel Coleridge, once wrote that the first great achievement of Greek philosophy was to extend the reach of philosophy itself, ‘not, indeed, as the product, but as the producing power — the productivity’. The difference here is slight and the application so profound as to be immeasurable, akin to enlarging humanity’s brain power, but the poet was using the word in the everyday sense. Had he been an economist, Coleridge would have said that the main achievement was not the product but ‘the producing power — less the effort and materials that were required to invent, learn and apply Greek philosophy — the productivity’.

Why does this distinction matter?

Because higher productivity is, in principle, a good policy objective, but its meaning has become lost in Australia’s public debate.

Many people who call for improvements in Australia’s productivity are really calling for more production, not higher productivity in an economic sense. Productivity in itself is neutral on the question of production levels. Australia can improve productivity and have lower, or the same, or higher levels of production. Productivity is not about production per se, but the ratio of output over input.

Australia could achieve higher production levels without any change in productivity by employing more people, i.e., by reducing the level of unemployment. If productivity improves faster than production, unemployment can increase, and at some point the social costs of high unemployment can outweigh the benefits from the productivity boost.

We could also increase production by making people work harder, i.e., by working longer and/or putting more energy into their work, but this would just increase the inputs, which won’t necessarily represent higher productivity.

Is the confusion over the concept widespread?

Yes. ‘Productivity’ now seems to be virtually an empty concept, or worse. There is much evidence to suggest that ‘productivity’ is now almost universally understood in Australia as a code word for forcing employees to work harder.

In a recent speech where the head of treasury called for a renewed emphasis on productivity, both the ABC and News Limited ran their reports under the headline ‘Australians must work harder’. This is not to blame the media outlets. In a sense they were reporting the story correctly, for this is now the way that the term is generally used and understood in Australia and, in an economic sense, abused and misunderstood.

Why should we aim to raise productivity anyway?

Because we can raise living standards by increasing output relative to input, whether by using the extra capacity this yields to increase production, i.e., to produce more or better goods and the associated income to increase consumption; or by realising the benefits in other ways, such as conserving scarce resources and protecting the environment without reducing output, or by enjoying more family, leisure and recreational time without reducing output — that is, by working less! As most of our waking lives are spent working in demanding jobs, any suggestion that this burden should be increased can only serve to undermine the purpose of raising productivity in the first place. In principle, working harder to raise productivity is a contradiction in terms.  

How do we actually measure productivity?

The most longstanding, widely accepted and useful way is by calculating ‘output per hour worked’, which is known as ‘labour productivity’. The convention is to divide the total value of production by the total hours worked.

There are problems that arise from the fact that we are dividing natural units (hours worked) into money values (GDP), but the index makes intuitive sense. The fewer hours that we need to work to produce any given value of output represents a productivity improvement. This is common sense, however difficult it is to precisely work out the ratio across a national economy over time.  

What about other methods?

The index known as ‘total-factor productivity’ (or ‘multi-factor productivity’) attempts to estimate the portion of output attributable to all the factors of production after deducting the value of any additional capital and labour inputs, and represents this as the rate of productivity growth.

The index doesn’t tell us what produced the extra (or residual) portion of output relative to the inputs. Many economists have described total-factor productivity as a ‘measure of ignorance’ or a ‘measure of error’. The index entails a long list of questionable assumptions, not the least of which are the assumptions that we can accurately measure the countless different forms of capital inputs, the difference made by returns to scale, and the separate contribution of the different factors when most production is a joint function of capital, labour and natural resources. Arguably this index creates more mysteries than it solves.

There are also many ‘partial productivity’ indicators that are used in particular industries, but the keyword here is ‘partial’ not ‘productivity’. 

So, we can only really measure the productivity of labour?

Yes and no. ‘Labour productivity’ is the best overall measure of productivity, but it doesn’t tell us anything about the causes of higher (or lower) labour productivity. The index simply distributes the value of the economy over one of the factors of production, albeit a very important one, but the result cannot be automatically attributed to labour. 

‘Labour productivity’ is just an elementary piece of arithmetic, not a direct measure of productivity or the expression of a causal relationship. We could divide the value of output by the number of churches or chairs just as easily. No-one would suggest that we have expressed a causal relationship, but the integrity of the sums would be identical.

To express the index’s aim simply: once upon a time a person would dig a ditch with a shovel; now that person can dig many more ditches in less time with a bulldozer. ‘Labour productivity’ will be much higher with the bulldozer, but most of the difference won’t be made by the labour, let alone by harder labour.

What use is the labour productivity index?

The main benefit is that it enables us to track trends over time against a consistent point of reference. The measure is particularly important at full employment, when there is no other sustainable way to increase production besides improving productivity. So long as productivity is rising, then production, consumption and the other benefits that speak of a rising standard of living can be realised without the economy hitting the wall and triggering inflation. 

For similar reasons, a farmer, for example, finds it useful to divide the value of output by the acreage in production. This won’t tell the farmer anything about the causes of the farm’s productivity, which will largely depend on what goes into and on top of the acreage and the conditions and methods of production and so on, but it will express the current level and the limits of the farm’s production without any change in productivity, and enable improvements to be traced over time. Once the whole acreage is in production, the only way to raise ouput is to increase the farm’s productivity.

The term ‘labour productivity’ is misleading if you interpret it as expressing a causal relationship. ‘Labour productivity’ is a disinterested proxy for gauging the productivity of the economy as a whole.

But won’t output go up ‘per hour worked’ if Australians work harder?

Yes, if everything else is equal, but — as already noted — this will not necessarily represent a real rise in productivity. Labour productivity is the best overall measure of productivity but it’s far from a perfect measure.

The index is also virtually useless for measuring productivity in service industries, for example, because outputs cannot simply be counted in the same way that cars can be counted in a vehicle-manufacturing plant. The value of the inputs, moreover, can generally only be measured by the quality of the output, which is difficult to measure at all.

Can you give more examples of problems with the index?

Other possibilities for measurement failure abound. Everything else being equal, labour productivity will go down when there are large capital investments that increase inputs before the output comes on stream (often noted in mining). By the same token, labour productivity will go up if companies run down their capital programs before the lack of investment causes real productivity to fall.

If the extra capacity is given over to protect the environment, this will not necessarily be registered as a rise in output, falsely implying no increase in productivity. By the same token, inputs that reduce the value of the environment might not be counted, giving a false measure of higher output.

Research and development will increase inputs without any immediate pay-offs in output, as will the diversion of labour inputs into education and training. By the same token, businesses can reduce inputs by cutting their research and training costs, which will reduce productivity in the medium and long terms.

There are also ways that firms can shift input costs onto customers (e.g., automated telephone answering systems), but this will not necessarily represent increased productivity for the society as a whole, since the extra input from consumers is not being counted. It’s important to appreciate that productivity is not the same as profitability, particularly in the immediate and short-term.

General economic conditions can also affect the index. During recessions, companies might retain unused capacity that does not really represent any technical decline in productivity; during booms, they might temporarily stretch capacity to unsustainable levels, implying a false rise in productivity. 

There are many reasons why the index can vary without reflecting what’s really going on with productivity, including straight out measurement errors. As the index is compiled from large aggregates of the economy and the percentage changes in growth from quarter to quarter and year to year are usually small fractions of 1 per cent, it’s reasonable to assume that the margin for error in the act of measuring is often larger than the differences being measured. All these issues arise with the other indexes, but they have more besides.

Evatt Q&A on ProductivityIn this light, how seriously should we take the index?

Small fluctuations in output per hour are best ignored and short-term variations are unreliable. What matters is the trend over the medium- and, certainly, the long-term. There is a well-known saying by the US economist, Paul Krugman, who once wrote that ‘Productivity isn’t everything, but in the long run it is almost everything’. The most important words here are ‘in the long run’.

What is Australia’s productivity record? 

Annual increases in labour productivity have averaged about 2 per cent a year since the late 1960s; sometimes a little lower, sometimes a little higher. There’s a celebrated period in the mid to late 1990s when the annual increase in labour productivity rose to about 3 per cent. It’s been around 1.5 per cent per annum in the new century. In the (latest) year to March 1913, it was about 2 per cent. 

What caused the rise in the mid to late 1990s?

No-one really knows, but there are many theories.

Neo-liberals always like to claim that it was the long delayed response to the reform program that began with the floating of the dollar in 1983 (privatisation, etc.), which has timing trouble on both sides of the story.

Others suggest that it was just a measurement error, i.e., the index was picking up an increase in work intensity. This would also explain why the work/family/life issue emerged in public debate over recent times, and why most people now think that higher productivity means ‘working harder’. It would also explain why the index went back to normal. Obviously you cannot increase work intensity indefinitely, and to the extent that the pressure to work harder is over the top, employees will invariably find ways to re-establish a sustainable equilibrium.

Others think that the absence of a recession made the figures look better, while many attribute the rise to the revolution in information and communications technology, with the gradual take-up reaching a critical mass that boosted productivity and then slowed as adoption stabilised, another explanation that would account for both sides of the story.

There could have also been a rare combination of reasons. Indeed, there is no inherent reason why there has to be a link between the causes of improvement from one measure to the next, or why many circumstances cannot occasionally converge to produce a ‘rogue poll’, so to speak.

So no-one really knows what improves productivity?

Yes and no. The invention and introduction of new technology is the most obvious and universally agreed cause of improvements. From the wheel, the lever and the screw, through to the steam engine and electricity, and on to semi-conductors and the internet, the level of investment in capital and labour-enhancing technology is the most obvious and universal difference between the productivity of traditional and modern societies, or rich and poor nations. The level of capital investment in the technological means of increasing the rate of production defines the frontier of contemporary levels of productivity.

The second most important and universal cause is the level of education and training, since this enables the best use and most rapid diffusion of technology, and ultimately plays a fundamental role in improving the rate with which new developments can occur in the first place. From literacy and numeracy through to using computers and on to learning by doing, education and training is a basic condition for the creation and development and especially the diffusion of more productive technology, processes and resource organisation.

But as to what causes what in any given period, the level of aggregation in the productivity index is so large and the capacity for human ingenuity so imponderable that it defies precise explanation and there are many mysteries.

Technological development can occur in jumps or incrementally. It can begin in small units and cause slow but large changes over time, or be introduced in big units to create large changes in short periods that then stop. A seemingly innocuous development at one level of the economy can open the way for large changes to another, and there can be any number and variety of results as the process unfolds in time. It took 70 years before sailing ships were overwhelmed by steamships, and another 30 years before the last of the sailing type left the merchant trades in the 1940s. Container ships were introduced in the 1950s, but did not represent a larger proportion of the world fleet than the conventional cargo ships until 2005. All the while, the productivity of ocean-going transport was slowly rising. Time matters, and matters in unpredictable ways.

The causes of higher productivity can usually be established for individual firms, but become more complex at the industry level, and often boil down to educated guesswork or crude ideological projection at the national and global level.

So, higher productivity is basically about developing policies that will encourage technological research and development and the diffusion thereof, mostly by investing in education and training?

These are the most obvious and universal conditions, but there are many other strands to an optimum policy menu, including a fair distribution of the benefits.

A striking coincidence is that the slowdown in the productivity growth of the mid-1990s began about two years after the enactment of the Workplace Relations Act in 1997, which is generally agreed to have unbalanced bargaining power in the workplace in favour of employers. Growth slowed even more about a year after the introduction of the notorious WorkChoices legislation at the end of 2005. Reducing input costs by reducing wages and conditions might raise production and profitability in the immediate and short-terms, but won’t represent any real improvement in productivity. Such policies are likely to reduce productivity growth in the medium and long terms by souring industrial relations. Why would anyone co-operate in raising productivity when all the benefits are going one way? Insofar as this leads to industrial conflict, production may also fall.

Yet there are many other conditions that can also have a bearing, about as many as can affect production itself. Adequate economic infrastructure is obviously important. So is social infrastructure. If a nation’s health system declines, so can productivity. Technology and education are crucial for helping to create more innovative products, processes and resource organisation, but so too is stability. People become better at what they do the more they do it. New employees often have to put more effort into their work but fail to reach parity at the start, and there might not be any improvement in their productivity before they have learned by their mistakes. High turnover will generally reduce productivity. An adequate social security net is important, for people will cling to jobs in which they have less aptitude if the alternative is catastrophic. The list goes on.

What’s the responsibility of trade unions in raising productivity?

In a direct sense, relatively minor. The main levers are in the hands of employers, while establishing generally favourable policies and conditions is the responsibility of government. Trade unions can be useful sources of advice on the likely benefits of particular proposals, since they look at the workplace from the perspective of the employees and bring a unique voice to the table. It is easy to forget that incremental practical experiment and random variation have probably made workplaces the most common source of technical innovation through history. At the same time, it’s not difficult to find workers who can tell of counter-productive measures introduced by management. Trade unions also have a vital role in striving to ensure that the benefits of productivity gains are fairly distributed and that any displaced labour is well looked after.

Why do employers and commentators often blame unions for impeding productivity growth?

Mainly for political and ideological reasons. There is a virulent anti-union lobby in Australia that will blame unions for any and everything. Employers also often portray unions as anti-productivity when they are really talking about taking away conditions to raise production, or refusing to raise wages to increase profitability, or as just part of the bargaining process. Or perhaps they want to deflect attention from their own responsibility for raising productivity.  

What questions should journalists ask employers who complain about low productivity?

In the first place, they should clarify whether they are talking about higher production or higher productivity, i.e., whether they are just seeking to increase output by increasing input or really trying to improve the ratio of output over input. If the latter, a whole host of potential questions arise. Have employers increased their training and research budgets? What plans do they have for investing in higher capital productivity? What is their track record in these areas? How do they envisage distributing the benefits? Have they consulted with their employees and unions on their plans? And so on.

What questions should they ask politicians?

Again, it is important to clarify whether they are using the term in an economic sense, or as just a euphemism for raising production or profitability. It’s also important not to accept that there is any kind of crisis just because of small variations in quarterly or annual estimates of productivity growth. It’s the medium- and particularly the long-term trend that matters.

Productivity is an interesting and ultimately very important issue, but it’s also complex, and there is a great deal of imprecision and theory associated with the concept. Public discussion and debate over productivity is, in general, a good thing, but it’s not an issue that politicians should be allowed to just glibly drop and run in media grabs and interviews to score political points, let alone turn into a mindless slogan. One of the things that we could all try to do for productivity in Australia is to rescue the concept from politics.

That said, the most obvious question to test politicians on their commitment to productivity at the moment is whether they support the so-called ‘Gonski reforms’ to improve school education. These reforms not only stand to help create a country that is more able to develop and adopt better products, processes and organisation, they will also help to create a fairer nation. It will cost a good deal of public money, of course, but more important than any product, it will raise the nation’s inherent ‘producing power’, as Coleridge might have said.


This Q&A has been produced by the Evatt Foundation in conjunction with ‘The Productivity Zombie’, a public forum hosted in association with Sydney Ideas at the Sydney Law School on 24 June 2013. The forum featured addresses by:

John Quiggin, Professor of Economics, University of Queensland, and author of Zombie economics: how dead ideas still walk among us (Black Inc, 2012);

Carmen Lawrence, Professor of Psychology, University of Western Australian, Panel Member of the Gonski Review of School Funding, and former Premier of Western Australia and former minister in the Keating Government;

Frank Stilwell, Emeritus Professor of Political Economy, University of Sydney and Executive member of the Evatt Foundation;

Christopher Sheil, Fellow in History at the University of New South Wales and President of the Evatt Foundation.   

Forward your questions, issues and stories about your own experience of ways and means of raising productivity to Evatt at: admin.evatt@sydney.edu.au  Copies of the printed booklet, Evatt Q&A on Productivity, are avaiiable from the Evatt Foundation.