What's productivity and how could AI affect it?

The Age

3 days ago

On the Sunday after Anthony Albanese's stunning 2025 election victory, Treasurer Jim Chalmers was given the job of publicly explaining the government's second term agenda. For three years, Chalmers and Albanese and the rest of the government had talked about dealing with inflation and cost-of-living pressures to the near exclusion of almost every other issue. But in that Sunday morning interview, the treasurer signalled a big change. 'The first term was primarily inflation ... the second term will be primarily productivity without forgetting inflation,' he told the ABC's Insiders program.
It was recognition that with price pressures finally easing, dealing with an issue that has been pivotal to the improvement in our living standards over the past two centuries was now front and centre. Three months later, Chalmers and Albanese and 23 experts from around the country will sit down in the federal cabinet room to discuss how to address the productivity challenge.
People often talk of productivity without knowing what it actually means. Employees tend to think it means working more hours; business owners think it's got something to do with reducing the costs of operation; politicians seem to think it's an economic Magic Pudding.
Since the global financial crisis in 2008-10, the rate of productivity growth has slowed around the world. In Australia's case, it is growing at its slowest pace in 60 years. Why? What is productivity? Can our lives improve without it?
What is productivity?
In an economic sense, productivity is a relatively simple concept – it is output divided by inputs. Let's use a widget-making factory as an example. For every 10 boxes of steel (input), the factory makes 1000 widgets (output). That equates to 100 widgets for every box of steel.
I buy equipment that reduces steel wastage during the production process, cutting my steel needs to nine boxes. I now produce 111.1 widgets for each box of steel. My productivity has soared by 11.1 per cent. This type of productivity is often called 'capital deepening'. It's a new machine or technology that allows us to boost output or reduce an input.
Productivity can also be increased by working smarter. I send my workers to Widget University to get a degree in widget making. They come up with a new way to produce widgets, making 1200 from just nine boxes of steel. This is an example of 'human capital'.
Mix capital deepening with human capital and you get what is called 'multifactor productivity'. It means how you combine new technology with your workforce to boost your output. Over the past century, more than half of all productivity in this country has been of the multifactor kind. The rest has been due to better technology.
By increasing our widget-making productivity, the overall price for our widgets falls. It enables a business to better pay their staff. And the falling price enables people to buy more. Those people have more income to spend on something else (and make it cheaper for a business that uses widgets as part of their production process).
There's also the concept of labour productivity – this effectively measures output per hour of labour. Over the past century, we've produced far, far more even though we spend less time at work.
It's not just widgets. Productivity is vital to the services sector, which accounts for the vast majority of our economic activity and employs more than 90 per cent of all Australians. One of the most creative parts of the services industry is music.
In the late 19th century, to be a music star required playing in front of an audience. Over the past century, technology – from wax cylinders to online streaming – has meant the entire world can hear the musician's output. Technology has meant that music itself has changed. Mozart needed an entire orchestra for The Magic Flute. Bob Dylan needed only six other men to upset the entire folk music scene by playing electric instruments at the Newport Folk Festival in 1965. And, 14 years later, Kate Bush used the first wireless headset microphone – setting a template for dancing and singing live performers from Madonna to Taylor Swift.
The graph below gives you an idea of how much labour productivity has changed since the start of the 20th century.
In 1901, an Australian needed to work 473 hours (at the average wage) to cover the cost of a new bicycle. By 2019, this had fallen to just six hours. Without productivity improvements, there would not have been such a huge fall in the relative cost of the humble bicycle.
Through the 20th century, the rate of growth in Australia was about 2 per cent a year. That rate meant the number of goods and services produced by Australians increased seven-fold while at the same time the hours worked by Australians fell substantially.
But productivity growth is slowing. In the 1960s, productivity in Australia was growing about 2.5 per cent a year. By the first decade of this century, it had slipped to about 1.4 per cent. And for the past 10 years, it's been closer to 1 per cent.
Does it really matter that productivity growth has slowed?
Just like compound interest, a small change in the rate of productivity growth can have long-term repercussions. The Productivity Commission says if productivity growth averages 2 per cent a year, economic output per person doubles in 35 years, triples in 55 years and increases five-fold over 85 years. But if productivity grows at 1 per cent a year, it would take 70 years for economic output to double.
You might be someone who doesn't measure their life by how much they produce. But you probably put a value on time not spent at work. According to the commission, Australians are selling themselves short on the 'leisure dividend' provided by productivity.
The 11.5 million people of Belgium have about the same GDP per person as Australia. But Belgium's productivity is higher than in Australia. If Australia's labour productivity was at Belgium's level, Australians could work four hours fewer a week without a reduction in income. If we reduced our working week by just an hour, we could boost our GDP per person by 25 per cent.
The commission says Australians have maintained our very high living standards over recent years by making up the productivity gap by working longer than comparable nations. Productivity is not just about being richer – it goes to our quality of life.
How important is the light bulb to productivity?
Let's illuminate this idea. People today live in a world of light. There's so much light, economists use photos from space to determine which countries are growing richer. But 160 years ago, the world was much, much darker. A key issue was not just the absence of electricity and lightbulbs but the sheer cost of producing light.
US economist William Nordhaus in the mid-1990s carried out research that went to the importance of cheap light. Nordhaus chopped wood and burned various types of candles to get a baseline on how much effort went into light creation. Chopping wood 10 hours a day for six days created enough fuel to create about 1000 lumen hours of light or about what you get from one light bulb in less than an hour. In productivity terms, the output is light divided by the input of wood and the 60 hours of labour needed to cut that wood.
If you spent your 60 hours creating tallow candles – which are made from the fat of animals – you would have enough candles to burn one a night for almost 2½ hours. Candles made from spermaceti, which is harvested from sperm whales, burned a little longer than tallow ones and did not give off the stench of burning animal. But they were more expensive (and the world hunted sperm whales almost to extinction). Oil lamps did better still, but were expensive and dangerous.
Then along came the electric lightbulb in 1879. By 1900, a single carbon filament bulb gave you 10 days of continuous bright light (without the smell of candles or the dangers of oil lamps) for the equivalent of a 60-hour week of work. Two decades later, bulbs delivered five months' light for the equivalent of that 60-hour week. And just before Nordhaus' research, bulbs were giving a decade's worth of light. Fluorescent bulbs give the same quality light for the equivalent of 52 years.
Nordhaus' work highlighted the collapse in the cost of light and the productivity enhancement of one of the world's most important inventions. Over 160 years, society went from stinking tallow candles that barely illuminated the pages of Twenty Thousand Leagues Under the Sea to a point where our towns and cities are never dark.
Why did abattoirs give us cheap cars?
Henry Ford is revered in business circles for the way his company brought cheap cars to the general population. Vital to his success was the moving conveyor belt. Rather than workers moving around a factory putting parts to a new car, a timed conveyor belt would bring the building blocks of a car to the staff. The idea of a conveyor belt was not new. Abattoirs brought carcasses to workers on belts, making it easier and quicker to carve up a cow or a sheep. Ford's breakthrough was to bring that concept to cars.
If you want evidence that productivity increases incomes and produces extra leisure time, then Ford is a prime example. Soon after he introduced his conveyor belt, staff started to leave because they were bored by the repetitious nature of the work (and also the pressure it put on them to complete their element of the vehicle). In response, Ford offered his staff $US5 a day as part of a profit-sharing arrangement. Critics said $US5 a day would bankrupt the company.
Instead, potential workers flocked to Detroit to work for Ford. Hours were also reduced, enabling the company to work three shifts a day. Productivity delivered higher pay, increased production and lower prices. A Model T Ford in 1908 cost $US825. In 1925, they were being sold at $US260.
What is slowing productivity growth?
While new technology and education have historically been the biggest drivers of productivity growth, they are not the only ways.
The World Bank, in a report in 2008 into Eastern Europe, noted that government policies can be just as important. Good governance, macroeconomic stability, competition, infrastructure quality, labour market flexibility, skill upgrading and financial deepening are all critical to productivity.
One of history's great economists, Adam Smith, noted in his book The Wealth of Nations that a lack of competition is not good for consumers.
'People of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some contrivance to raise prices,' he wrote back in 1776.
Smith was writing just as the Industrial Revolution was starting. It would unleash the productivity surge that delivered improvements to living standards around the globe that today we take for granted.
But Smith's concern was that men of trade would go out of their way to stop active competition. Competition – or the lack of it – has been identified among some experts as one of the reasons why productivity has slowed over recent years.
Rod Sims, a former boss of the Australian Competition and Consumer Commission (ACCC), has argued that from beer to airlines to banks, Australia simply doesn't have enough competition to force companies to find ways to be more productive. 'If Australia had fewer markets run by dominant oligopolies, and more benefiting from strong competition, we would have less inequality and higher wages and productivity,' he wrote.
Unions, in their push to protect wages and conditions for their members, can also impede productivity. Ahead of the economic roundtable, the ACTU proposed a shift to a four-day working week on the proviso that a person's pay did not fall despite their shorter hours.
Loading
The Australian Chamber of Commerce and Industry recently complained that the Maritime Union of Australia has gone out of its way to oppose automation systems across the nation's ports. This type of automation is normal in overseas ports. But, said chamber chief executive Andrew McKellar, the MUA stands in the way of technology that might reduce the number of port workers but which would save consumers time and money. A World Bank report looking at the world's best ports ranked Australia's five ports in the bottom 30 per cent, with Adelaide the highest at No. 264.
One of the loudest voices calling for regulation of artificial intelligence has been the ACTU, which has effectively argued for unions to control the shape of new tech laws. The Productivity Commission argues that a light regulatory touch, including the use of existing laws, to deal with fears about AI may pave the way to a $116 billion boost to the economy.
What else is slowing productivity?
In 2017, the Productivity Commission released its Shifting the Dial report, which outlined changes the country should examine to make it richer and more productive. The dust on that report is a mile high, just as the commission starts a new report into the same issues. It's a similar story to the commission's 2023 report Advancing Prosperity, which built on some of the Shifting the Dial 's proposals and then added pages of new ideas.
Why are these reports dust? Because the answers proposed by the commission are politically fraught.
Shifting the Dial 's first recommendation was to free up resources in our local health systems and hospitals. But that would immediately get in the way of the states, who have a say over these areas.
The second proposal was to create a 'do not do' list on low-value health interventions. Instead of precious health dollars going into procedures that provide a marginal health benefit, it would go towards areas that give better value. Such a proposal would deliver an income hit to GPs and specialists – two very powerful and influential lobbies.
Advancing Prosperity 's opening pages advocated reforms to higher education, including new funding models and an overhaul in the pricing of certain courses. This just two years after the Morrison government's own contentious changes had been put in place. Its proposals to reduce industrial relations regulations emerged just months after the Albanese government tightened the very same regulations.
Both reports promised big increases in economic growth. But the political pain from such proposals could be measured in seats in a parliament (both federal and state). And the pain felt by political parties would be due to the way the voting public reacted to proposals that would definitely cause economic turmoil for some parts of the population.
Aren't there downsides to productivity changes?
Two areas of potentially large jumps in productivity are artificial intelligence (AI) and automation. Both tap into fears about technology and productivity that have been evident since the Industrial Revolution. The Luddites of the 1810s smashed textile machines, fearful that skilled weavers such as themselves would be put out of work by technology.
The argument today against automation and AI is similar to those heard in the pubs around Nottingham in the 1810s. AI has the added spice that it will lead us to a world inhabited by T-800s (fans of Terminator will understand). The Productivity Commission estimates AI alone could deliver a $116 billion boost to the nation's economy over the next decade as mundane work activities are completed quickly and competently by ChatGPT or Gemini while humans get on with more valuable work.
Better use of our personal data, if we agree to share it with the right companies, would be worth an extra $10 billion a year. But the same report kicked off concerns among writers, singers and anyone with a creative bent after it suggested AI freely use Australian-created content without even a couple of dollars to compensate for their work.
In Australia, mining truck drivers have been replaced by automated vehicles over the past 10 years, yet the number of people working in the mining sector continues to grow – this year, a record 302,000 people worked the nation's mines.
The Luddites' concern that they would be put out of work was well founded. Machines did replace the groups of well-paid weavers. If you were a lifelong weaver whose family depended on your weaving income to survive, the advent of weaving machines was akin to the arrival of a T-800. But the community benefited from the reduced cost of cloth. Some of that reduced cost flowed to the owners of the machines, some of it flowed to the workers they employed, and some of it flowed as a benefit to consumers who were able to buy cheaper products.
At the start of this century, the Productivity Commission estimates about 84,000 people were employed globally in the video rental store Blockbuster. But Blockbuster is gone – replaced by Netflix and other streaming services that employ about 11,600 people while offering far, far more than even the world's largest video rental store.
Without that technical breakthrough, all people would be poorer – both in terms of income (especially what they could buy with that income) and in the amount of time spent at work.
In early August alone, both the Commonwealth Bank and Atlassian announced job cuts, with people replaced by AI.
Productivity improvements – particularly those borne from technological advancements – do lead to the end of jobs. From whalers to chimney sweeps to manufacturing, better ways to make things or deliver services lead to people losing work. Anyone denying that is having you on.
Even the commission admitted that there would be people left behind by the technological solution, suggesting some would ultimately rely on unemployment benefits. But the same technological changes do boost the overall number of people in work, creating new jobs, occupations or entire industries.
And this is the problem with new forms of technology. Boosters promise a world of rainbows and sunshine. Supporters of AI, for instance, have talked about annual increases in GDP of 20 per cent or more. That's double the rate that China enjoyed at the start of this decade, which itself was unprecedented (and which, ultimately, slowed). And there's no explanation of how AI itself, rather than the huge data centres built to support the technology, will generate growth rates that have literally never been experienced anywhere.
Loading
Where else might productivity come from?
Automation and AI offer opportunities for productivity growth, but perhaps the biggest change might come from a medical breakthrough. People have benefited from huge changes in medicine over the past two centuries. We are much less likely to die at birth, during our early years, in accidents and from disease. Our food and water are safer.
In our middle and elderly years, our quality of life is far superior to what it was even a few decades ago. When the age pension was introduced in 1909, the average life expectancy for a man was 58 years. Today, it is 82 for men and 85 for women.
If researchers find a way to cure cancer or dementia (two of the leading causes of death in Australia), then not only will our collective quality of life improve but there will be a lift in our overall productivity.
Another area of possible improvement is through the better use of data. Businesses and governments now collect vast amounts of information. That information could be used to deliver insights that will improve productivity.
Australian statistician Henry Lancaster used data from several national censuses to link an outbreak of rubella that hit the country in 1898-99 with a group of people described as 'deaf-mutes'. It was the first time the link between rubella and health problems in unborn children had been established. In 2018, the World Health Organisation verified that rubella had been eradicated from Australia.
AI proponents, for instance, believe their technology will super-charge the pace at which data is processed in the search for patterns like that found by Lancaster. His work took years – AI promises success in hours.
But there are also concerns that the big productivity breakthroughs due to technology are over. US economist Robert Gordon, who has described himself as the 'prophet of pessimism', argues that 'big-bang' productivity enhancements have effectively come to an end. The lightbulb, a reliable internal combustion engine and the telephone were created in a three-year period around 1876. Other changes, such as indoor plumbing and refrigeration, were invented not long after.
Even the smartphone is dwarfed by previous advances. We need to get used to only moderate productivity improvements because game-changing advances are unlikely. In public appearances, Gordon displays a photo of a smartphone and a toilet and asks, 'Which would you rather give up?' Gordon admits AI will contribute to a lift in productivity growth, but not in the same way past technological changes have upended the economy. AI, he argues, will not replace as many jobs as feared.
Another thesis is that, as the global population ages, dynamism in business or technology slows. There are also concerns that the fall in global interest rates over the past 30 years has seen money flow not into entrepreneurs changing the face of our daily lives, but into a property boom that's simply inflated the price of bricks and mortar. Housing is important. But a breakthrough in splash-backs is unlikely to boost our collective productivity levels.
Whatever the future – AI driving our cars or beautiful tiles in our bathrooms – productivity growth will be pivotal to our living standards forever more.