Stop worrying about productivity. We’re doing okay.

Australia’s productivity ‘crisis’ is not what it seems. We’re just not measuring it properly.

The most often-quoted line on this subject is from Paul Krugman, the US economist. “Productivity isn’t everything, but in the long run it’s almost everything,” he said. “A country’s ability to improve its standard of living over time depends almost entirely on its ability to raise its output per worker.”

But unless we measure it properly – and we’re not – fundamental decisions about investment, government budgets and wage levels are likely to be wrong. Official statistics are always, and inescapably, rubbery but we accept them with more trust and fewer questions than we apply to holy writ.

Even the census is rubbery, out of date even before the first results are released. Most of the other sorts of data are based on estimates or surveys. The statisticians are honest about the rubberiness, often giving estimates of how far out they may be: but the rest of us take them as revealed truth.

Productivity means producing more for a given input. It’s how the economy is able to afford a rising standard of living, how we can maintain better and more expensive services. In turn, that’s helped to produce a healthier, better educated population that also lives longer.

Productivity allows employers, and the economy at large, to afford higher real wages. Without that, economically sustainable wage rises would be limited to making up for inflation. Your purchasing power would never move.

So if productivity is under-estimated – and there’s strong evidence that it is, particularly in industries which employ the most people – employees, their families, and eventually everyone, suffer.

The basics

First, we need to understand what productivity is. There are three main types.

The most important is labour productivity, which measures the value of output per unit of labour. This is what’s usually meant by ‘productivity’ in public discussion.

Then there’s capital productivity, which attempts to quantify how output compares with the amount of capital invested.

And multifactor productivity attempts to put the two together to produce an overall picture.

There’s no attempt to measure the efficiency and sustainability with which we use natural resources, which are the third and most fundamental factor of production. So the productivity measures assume that we can go on increasing output forever without worrying about the limits of the natural world.

Estimates of productivity growth in most developed countries have been declining for some years. The figures bounce around a bit, but the trendline tells the story of the past three decades.

Australia is not alone. An apparent decline in productivity has bedevilled most developed nations and, as this chart shows, Australia isn’t doing all that badly by comparison.

But is all this as dire as it appears?

The madness in the method

Attempts to measure productivity go back at least 250 years but it wasn’t until the middle of last century that it became a major factor in public and government considerations. In the 1960s, the US Bureau of Economic Analysis was asked to construct a comprehensive method of measuring productivity across the entire national economy. The concepts being used today have refined, but not fundamentally changed, those of the 1960s.

But even then, the productivity method was problematic and probably inaccurate when allied to the majority of industries. The system works reasonably well in the production of goods. A widget-making factory which produced 1000 widgets last year and 1020 this year with the same labour increased its productivity by 2%.

It works because output can be reliably calculated: the number of widgets produced. But how do you calculate the output of a hospital, a school or a police force?

For services, and particularly for those services which are dominated by the public sector, the conventional system doesn’t work very well. Over the past half-century, the situation has become progressively worse: the goods sector today accounts for a lower proportion of economic output than it did 50 years ago. And the services sector – particularly the public sector industries, healthcare, education and public administration – have become more important.

The shift from manufacturing to services as a share of the economy has been going on for longer than you may have thought. Even from the earliest days of European settlement, Australia’s service sector has (except for a period in the early 19^th century) massively outstripped even agriculture. The decline of manufacturing, paradoxically, began at the end of World War 2: even the birth of a car industry did not save it. At the same time, the service sector boomed and kept on booming.

This graph also puts our various mining booms into perspective. They all had their effect but only the Gold Rush has been truly transformative – and that didn’t last.

It's a similar story in employment.

The services sector – that is, everything other than those three industries – has increased its relative contribution, but the picture is complicated.

All the relative growth over the past half-century has been in the three industries which comprise the non-market sector – healthcare and social assistance, education and training, and public administration and safety. Those three are dominated by the public sector, and they’re very hard indeed to measure accurately. This table shows in more detail how the economy has shifted over the past 50 years. The proportion of total output being produced by the goods sector – that is, agriculture, mining and manufacturing – has declined. That’s mostly due to the deterioration of manufacturing.

Here's a more nuanced view. Over the past 40 years, the industries in which most people work have changed dramatically, both in the number of people they employ and in the value they produce.

The most profound changes are in manufacturing, and in healthcare and social assistance. Manufacturing now employs 25% fewer workers than in 1984, but the workforce in healthcare and social assistance has soared by 318%.

Agriculture (which includes forestry and fishing) has reduced its share of national employment by 4.1% over the period but increased the value of its output by 0.9%. So its overall labour productivity has gone up.

The share of Australian workers employed in mining has barely moved (up 0.7%) but its share of national output has soared (up 4.9%). Its labour productivity has therefore also risen.

Manufacturing’s share of employment has fallen more sharply (down 10.7%) than its share of output (down 7.7%) so, although manufacturing has been in serious trouble, its labour productivity has improved significantly.

But look at healthcare and social assistance. Its share of output has increased too, by 3.3% – second only to mining. But its share of national employment has risen by a whopping 7.3%, by far the highest of any industry. So, although healthcare and social assistance is more important to Australia’s economy and society now than it was 40 years ago, its labour productivity figures – at least in the way they’re officially calculated – are sharply down.

Over the last 30 years, overall labour productivity has been in trouble. The average annual increase (for all industries) was 1.39% in the decade from 1996 to 2005, 1.22% from 2006 to 2015 and only 0.43% in the most recent decade.

As the following chart shows, it was very unevenly spread. Agriculture was hit by drought and gyrations in world prices, but recorded very strong growth. So did mining, though it also has been affected by lower commodity prices.

Information media and telecommunications has benefited spectacularly from the internet age. Administrative services, which include call centres, travel agents and employment services, have also done well.

But services dominated by the public sector – public administration and safety, education and training, and healthcare and social assistance, have lagged badly. These are necessarily labour-intensive and together employ around 4.8 million people ranging from doctors and nurses, teachers, public servants to police, firefighters and judges.

Calculating labour productivity in these fields is far from straightforward. And the evidence is that the task is not being done well.

Research in Australia and overseas shows healthcare and education, in particular, are much more efficient than the official figures claim – and, therefore, people working in those professions are likely to be paid less than they should be.

Measuring health

Healthcare productivity is conventionally measured by the price of a service compared to a price index, but it ignores what that service can achieve. Cost matters; quality doesn’t. If it costs more but saves your life, productivity is said to fall, compared with a cheaper service that lets you die.

The official figures from the Australian Bureau of Statistics show that over the past decade, labour productivity in healthcare and social assistance fell by an average of 0.28% a year. But the conventional measure – gross value added per hour worked – turns out to be highly misleading. By this method, the only way productivity can be improved is if costs go down.

But, as it turns out, quality matters more than cost. Ignoring changes in the quality of healthcare means that only the cost of new methods and new technology is measured, not the benefits.

An analysis by the Brookings Institution found:

“In the traditional approach, the good is the health care service or good actually purchased: a doctor’s appointment, a hospital stay, a prescription. But … these purchases are better viewed as intermediate inputs into the production of what the consumer truly wants – better health.”

The Brookings researchers found that an index of price inflation in heart treatment was +4.7% according to conventional measures. But when quality was taken into account, the index was minus 14.4%: the real cost-to-benefit ratio has improved dramatically:

“We find that the quality-adjusted price of heart attack treatment fell by about 14 percent per year between 1984 and 1994.”

Australia’s Productivity Commission has found similar results. Using methods developed in the United States, they found that rather than going backwards, healthcare productivity was strongly positive.

“Our research suggests that parts of the healthcare sector have experienced robust growth,” they wrote.

“Quality-adjusted multifactor productivity grew by about 3% per year between 2011-12 and 2017-18 for the subset of the sector we studied, healthcare used to treat: cancers, cardiovascular diseases, blood and metabolic disorders, endocrine disorders and kidney and urinary diseases. Together, treatment of these diseases accounts for around one third of healthcare spending. To put this in perspective, multifactor productivity growth in the market sector was estimated to be around 0.8% per year over the same period.

“Quality improvements, not cost reductions, were the big contributors to productivity growth. Health practitioners’ ability to understand, diagnose and prescribe has been transformed by medical advances. These quality improvements have made us vastly better off. Indeed, the biggest contribution to productivity growth has come from advances in saving lives.”

Measuring education

Very similar shortcomings have been found in education. An example: when a school employs more teachers and reduces the size of classes, labour productivity goes down. For the purposes of these figures, there’s no difference between a highly-skilled and effective teacher and an inept drongo. If the drongo is paid less than the competent teacher, productivity improves. And if a disadvantaged school gets the extra resources it must have to turn kids’ lives around, its labour productivity suffers.

This is what happens when you look only at cost and ignore benefit. Quality matters.

Britain’s Office of National Statistics, unlike most other agencies, measures the quality of education as well as its costs. The difference is substantial. Over the period from 1997 to 2019, the average annual growth in the productivity of Britain’s education system was 1.6% when quality was taken into account but only 0.5% when it was not. This is revealed even more clearly when we track the growing disparity over time.

Quality isn’t the only thing being ignored. The misclassification of cost inputs is also important.

Is educating a child through school and university only a cost, or is it also an investment? There is a massive body of evidence showing that an educated workforce is essential for economic growth and prosperity. It’s also, not incidentally, essential for a properly functioning democracy.

Current methods may take into account the increased earning capacity of university graduates, but that has little to do with the overall value of investing in human (rather than only physical) capital. Taking a broader and more realistic view would substantially increase not only the level of education productivity but also, probably, its growth rate.

One important study has found that the rate of return on investment in human capital through education in European countries averaged around 13%.

The failure to measure productivity properly produces some strange results, as British researchers have documented:

“Evidence suggests that, in OECD countries, the correlation between average educational attainment and ‘multi-factor productivity’ – the overall efficiency with which labour and capital inputs are used together in the production process – is negative (and statistically significant). Counterintuitively, this would suggest that the more a person is educated or trained, the less productive they become.”

What about the others?

The relationship of human capital investment and quality adjustment are under-researched in the public sector but almost totally ignored in market sector service industries.

In theory, the conventional measures of productivity ought to work in all market sector industries because producers are paid by their customers. So input and output can both be measured in dollar terms and compared. But is it as simple as that?

Price is reliable as a measure of worth only in a balanced market, in which buyer and seller have comparable power and information, and there’s the right amount of competition among both sellers and buyers. In the real world, those markets are rare.

Many factors can affect the price being paid for a product. The same good or service may attract a higher price if there’s less competition. Buyers may not know much about what they’re getting. Mistaking price for value can produce strange results.

When Rex and Bonza airlines began flying the capital city routes, the big airlines dropped their fares below a point at which the new entrants could compete. The economy became more efficient but the airline industry’s productivity fell.

When Qantas and Virgin succeeded in sending their competitors broke, they raised their fares massively. The economy became less efficient but productivity improved.

Productivity Commission research has pointed to other anomalies. Improvements in the supply of utilities (such as putting power cables underground and upgrading sewerage and waste disposal) have contributed to lower productivity figures because only costs, not wider benefits, are taken into account.

Over 65% of the output in the finance and insurance industries is imputed in a way that conflates production with the premium banks get for an above-zero level of risk.

Changes in retail and wholesale trading, such as longer trading hours or a greater range of goods on offer may not be fully captured in the retail price.

Intangible investments – that’s any investment, such as human capital, that’s not in physical goods – tend to be ignored or misallocated.

“Many activities that require a fixed investment for future payoff – such as staff training, brand development and organisational practices – are considered current (rather than capital) expenses for the purposes of the national accounts,” the Productivity Commission researchers wrote.

“Australian and overseas research has confirmed that broader definitions of intangible investment mean that the share of intangible assets is higher than is implied by the national accounts, and that its increase over time is more pronounced.”

The upshot

On the basis of the evidence we have, Australia and most other countries are miscalculating productivity in their economies. Most of this miscalculation has resulted in probable under-estimations which, when put together, have profound implications for the policy and, not incidentally, for wage growth.

As the economy has moved from goods and towards services, the problems associated with measuring service productivity have become progressively more acute. Measuring economic growth underpins most planning by governments and businesses, but it depends on the accuracy of the figures. If those data are incorrect, wrong decisions are likely to be made.

Low wage growth has dogged Australia for many years. Along with price inflation, it’s a principle cause of the current cost-of-living issues. As we’ve seen, a fair distribution of economic benefits between the owners of labour and capital requires wages to compensate for inflation and for productivity improvements.

Even using the current official data, wages have persistently failed to keep up with productivity. The real situation is even worse than it appears.

If employees do not get their fair share of the benefits of growth, the owners of capital get it instead. It’s not the only cause of income inequality, but it’s a substantial one.