Speech Why Productivity Matters

Photograph of Michael Plumb

Australian Business Economists Annual Forecasting Conference
Sydney –

Introduction

Thank you for the opportunity to speak here today at the Australian Business Economists’ Annual Forecasting Conference. There has been lots of discussion about productivity in recent years. In some economies this discussion has been about subdued growth in overall productivity, including in Australia since just before the pandemic. There has also been discussion about the outlook for productivity. For example, the extent to which artificial intelligence, quantum computing and other technologies will support future productivity growth. These are important issues that I expect will come up in discussions today.

In my remarks I’m going to focus on a different question: why does productivity matter? At the central bank we’re not experts in how to improve productivity. But trends in productivity are very important for the macroeconomy. In the context of the Australian economy, I will discuss how stronger productivity growth can support growth in aggregate supply, incomes and aggregate demand. I will then spend some time discussing recent productivity outcomes in Australia and how we’ve been thinking about those in our assessment of economic conditions.

But first, what is productivity? When we talk about productivity, we’re talking about how much output we get relative to what we put in. At an individual level, I increase my own productivity by making a shopping list before I buy groceries, so I don’t forget anything and avoid multiple trips to the supermarket. At the firm level, productivity might be improved by implementing customer relationship management software to streamline communication with clients and automate routine tasks. At the economy-wide level – which is what matters for the central bank and our dual mandate of full employment and low and stable inflation – productivity reflects a multitude of decisions like these. Ultimately it’s about how efficiently capital and labour are employed across the economy to produce goods and services.

How do we measure productivity? Economists typically focus on two measures: labour productivity, which measures how much output is produced for every hour worked; and multifactor productivity (MFP), which reflects how efficiently all inputs to production – such as labour, capital, energy and raw materials – are combined to produce output.

In a simple production function framework where a firm produces output using two inputs – labour and capital – labour productivity depends on two things. The first is how much capital each person has to work with. Providing workers with more or better capital – like machines or faster computers – can increase the amount of output each worker produces. This is referred to as ‘capital deepening’. The second is MFP. Improving MFP involves finding new ways to combine labour and capital to produce more output. For example, by reorganising a production line or using GPS technology to precisely guide machinery for planting, fertilising and harvesting. In this respect, labour productivity is not just about labour efficiency; it depends on firms’ decisions about how much capital to employ and how efficiently labour and capital work together to produce output.

In thinking about the relationship between productivity and aggregate supply, incomes and demand, I will focus mainly on labour productivity. This is because labour productivity most closely aligns with measures of economic living standards. It’s also easier to measure than MFP.

As you might sense, productivity is not about working harder, but working smarter. Many of the biggest productivity improvements have come from things that have made our lives easier, like computers, robots, the internet and smartphones – though personally I’m still questioning whether smartphones are productivity enhancing or a productivity sapping distraction.

Economists talk about productivity a lot. So I’ll now turn to the question of why productivity matters.

Productivity and supply

If productivity increases, the economy can produce more goods and services from all the available economic inputs. As such, productivity is a key driver of growth in the supply capacity of the economy, or potential output.

Productivity in Australia has been volatile in recent years but, looking through the volatility, is around the same level as in the few years before the pandemic. Productivity growth has also been consistently below the RBA’s projections for some time now (Graph 1). This has generated internal discussions about what trend labour productivity growth might look like in the period ahead, and what that means for estimates of potential output growth over the forecast period. The current assumption is that annual labour productivity growth will pick up to around one per cent in the medium term, which is close to its longer run average. This could be consistent with, for example, the rapid adoption of technology across many industries leading to higher productivity outcomes. However, the projected pick-up in productivity growth has not materialised in recent years and staff are currently assessing whether weak productivity outcomes are likely to persist.

Graph 1
Graph 1: Productivity - A one-panel line graph showing an index of productivity growth since 2015 against index forecasts from the November SMP of each year between 2015 and 2024 inclusive. The index shows that the level of productivity remains around the same level as it was in 2016, and that actual productivity growth has been significantly weaker than expected in 2021, 2022 and 2023.

Weak productivity growth in recent years has contributed to slower growth in the supply capacity, or potential output, of the economy than otherwise. Graph 2 shows one of our estimates of potential output, which is based on actual productivity outcomes observed in the data. The graph also shows a counterfactual path where productivity growth in recent years was higher, at its average rate in the two decades prior to the pandemic. This suggests that the size of the economy is a lot smaller than it would have been, had productivity growth been more like in the past (all else equal).

It’s important to keep in mind that, in this counterfactual world where supply capacity was much higher, incomes and demand would also have been higher too. Let me turn to that now.

Graph 2
Graph 2: Potential output - A one-panel line graph showing counterfactual potential output and potential ouput, both in chain volume billions of dollars. The two lines diverge from 2020 as counterfactual potential output growth is higher than potential output growth from 2020. This is the result if multifactor productivity growth had been growing at its pre-pandemic long-run (two decade) average.

Productivity, incomes and wages

While productivity growth contributes to growth in the supply capacity of the economy, it also contributes to growth in incomes and demand.

At times, labour productivity (output per hour worked) and real income per hour track one another closely (Graph 3). Looking through the volatility, both are currently around similar levels as in the period prior to the pandemic.

Graph 3
Graph 3: Productivity, income and Terms of Trade - A two-panel line graph from 1987 to 2024, indexed to September 2012. The top panel shows that labour productivity and real net national disposable income per hour tend to increase together over time, though at times there are divergences. The bottom panel shows the terms of trade relationship, which is volatile but increased substantially as a result of the mining boom (from mid-2000s). Together, these two panels show that real net national disposable income per hour grows faster than labour productivity when the terms of trade is increasing, as it did during the mid-to-late 2000s.

Other factors besides productivity can affect growth in incomes per hour. For example, higher prices for Australian exports can generate higher incomes domestically. So the terms of trade – the prices we receive for our exports relative to the prices we pay for our imports – can also be an important driver of incomes in the domestic economy. We can see this in the decade from the early 2000s: despite the slowing in productivity growth, real incomes per hour continued to increase, partly owing to substantial increases in the prices received for Australian exports like iron ore and coal. The surge in demand for our exports, particularly from China, supported profits in the mining industry and related parts of the Australian economy, as well as demand for labour and wages growth.

Over the longer run, labour productivity and real wages – as measured by average earnings from the national accounts – also tend to move together (Graph 4).[1] Over the inflation targeting period, labour productivity has grown at an average annual rate of 1.1 per cent and real labour earnings have grown at 0.9 per cent.[2] So, higher productivity not only benefits firms, it also benefits workers by increasing their purchasing power. The Productivity Commission has previously pointed to the productivity of bakers as a reason we can consume more bread or spend that extra money elsewhere – in 1901 it took 18 minutes of the average worker’s time to afford a loaf of bread, while today it’s just 4 minutes.[3] There must be a joke in there somewhere about how we spend our dough.

In the short run, however, growth in real wages and labour productivity can and do diverge as the economy adjusts to economic shocks. For example, and as noted previously, increases in the prices received for Australian exports can have an impact on domestic profits and wages (and without an increase in labour productivity). Ultimately, however, it is very hard for an economy to support real wages growth in the longer run without productivity growth.

Graph 4
Graph 4: Real wages and Productivity - A one-panel line graph from 1987 to 2024, indexed to September 2012, showing real average hourly earnings and labour productivity, both increasing over time. In the long run, the relationship is strong, though there are periods where real average hourly earnings is growing faster or slower than labour productivity.

Productivity and consumption

Productivity growth also tends to support consumption growth. When productivity and incomes are growing more strongly, people are able to spend more and consumption grows more quickly. Weak growth in consumption per capita over recent years has coincided with weak growth in productivity, real incomes and real wages (Graph 5).[4]

Graph 5
Graph 5: Productivity and Consumption Growth - A bar chart showing annual growth rates of productivity and consumption per capita, averaged over productivity cycles. The graph shows that labour productivity growth and real consumption per capita growth tend to move together closely over time.

Similar patterns have been evident in other economies, where subdued productivity growth has been associated with slower growth in household incomes and consumption (Graph 6). The exception is the United States, where growth in both productivity and consumption has been relatively strong.

Graph 6
Graph 6: Consumption Per Capita and Productivity - A six-panel line graph from 2012 to 2024 that shows consumption per capita and productivity for Australia, Canada, euro area, New Zealand, United Kingdom and the United States. For each country, consumption and productivity move together over time, apart from some very short-term deviations. Each country has experienced declining consumption per capita and labour productivty over the past couple of years, except for the United States, where both consumption and productivity have been growing strongly in recent years.

Recent trends in productivity

So far I’ve focused on the importance of productivity growth for aggregate supply, incomes and demand over the longer run. I’ll now turn to recent trends in productivity growth in Australia and some potential implications for the near-term economic outlook.

Discerning recent trends in productivity is difficult because of volatility in the data associated with the pandemic and other supply disruptions. Looking through the volatility, labour productivity growth has been low, averaging 0.2 per cent per year between 2017/18 and 2023/24 (Graph 7).

Graph 7
Graph 7: Productivity - A two-panel line graph showing, from 1985 to 2024, labour productivity in the top panel and multifactor productivity in the bottom panel. Trend lines are drawn through each panel for each productivity cycle, such that it shows labour productivity and multifactor productivity have slowed in recent productivity cycles, relative to the 1980s, 1990s and 2000s periods.

Reverting to the simple production function framework that I noted earlier, the slow growth in labour productivity over recent years has reflected slow growth in both MFP and the amount of capital available to each worker.

MFP growth averaged 0.2 per cent per year between 2017/18 and 2023/24, which was well below its historical average. Some have argued that slower MFP growth could reflect temporary factors. For example, tight labour market conditions over recent years have been associated with large numbers of individuals entering the workforce or changing jobs; this may have weighed on productivity as some individuals were trained or retrained and some firms adapted production processes to accommodate strong employment growth. If this was the case, MFP growth could pick up as the economy adjusts. However, work by some RBA staff finds that temporary factors like these have not been the primary cause of slow MFP growth, suggesting that structural factors could be weighing on productivity growth.[5]

Slow growth in the amount of capital available for each worker in the Australian economy – or a lack of ‘capital deepening’ – has also contributed to slow growth in labour productivity (Graph 8). Capital per worker was broadly unchanged for around five years leading up to the pandemic and – looking through the volatility in the data during the pandemic – is currently a bit below those levels. In other words, overall investment has not kept pace with the strong growth in employment recently.

Graph 8
Graph 8: Labour Productivity and Capital Deepening - A one-panel line graph showing labour productivity and capital deepening (the capital-to-labour ratio). The two lines are indexed to the December quarter 2019 (which is set equal to 100). It shows that the capital-to-labour ratio and labour productivity move together over time, and that both have been volatile over the past few years. Productivity growth in recent quarters has been supported by the recovery in the capital-to-labour ratio, though productivity is around 2016 levels and the capital-to-labour ratio is still below 2016 levels.

To help understand the recent slow growth in productivity, I’ll look at productivity outcomes in various parts of the economy.

I’ll start with the non-market sector – which includes the health care, education and public administration industries – where employment growth has been very strong over recent years. The level of measured productivity in some parts of the non-market sector is low relative to the aggregate economy. So, as the non-market sector has become a larger share of the economy in recent years, this has weighed on overall productivity growth in the economy. Our estimates suggest that the rising share of non-market employment lowered the economy-wide measure of labour productivity growth by around 0.3 percentage points per year on average from 2017/18 to 2023/24, as shown by the yellow bars in Graph 9.[6] This compares with around 0.15 percentage points per year over the previous decade, and so the recent effects have been a bit larger than in the past.

But there is more to the story about productivity and the non-market sector. I have emphasised measured productivity because it is very difficult to measure output – and therefore productivity – in parts of the non-market sector. The central measurement problem is a lack of meaningful prices for some non-market output, such as public hospital services provided to public patients.[7] This makes it very difficult to accurately identify quantities of output, which are needed to measure productivity. For example, research by the Productivity Commission suggests that productivity in the health care industry is higher than official estimates.[8] As such, the drag on productivity from the non-market sector may be overstated.

Graph 9
Graph 9: Contributions to Labour Productivity Growth - A bar graph from 2009/10 showing three different types of contributions to labour productivity growth: from growth in market sector productivity; growth in non-market sector productivity; and the contribution from reallocation between the market and non-market sectors. The graph shows that usually growth in market and non-market sector productivity is the source of total productivity growth, whereas the reallocation effect typically drags down total labour.

Noting the challenges of measuring productivity in the non-market sector, what’s been going on in the rest of the economy? Labour productivity growth in the market sector averaged around 0.6 per cent per year from 2017/18 to 2023/24 – below its average of 1.6 per cent over the previous two decades – though it picked up in 2023/24.

Table 1: Growth in Labour Productivity
Average annual growth rates (per cent)(a)
Sector 1998/99 to 2017/18 2017/18 to 2023/24
All industries 1.3 0.2
Non-farm 1.1 0.1
Market sector 1.6 0.6
Market sector ex mining 1.4 1.0

(a) Average growth rates calculated between financial years.

Sources: ABS; RBA.

While the level of productivity in the mining industry in Australia is higher than in other industries, productivity growth in that industry has declined over recent years. Excluding mining, productivity growth in the market sector since 2017/18 has averaged 1 per cent per year, though this is still lower than its average over the preceding two decades and well below the rates recorded during the high productivity growth period in the 1990s.[9]

More generally, a range of explanations have been provided for the slowing in productivity growth globally since the 1990s. A well-documented one for Australia is declining ‘economic dynamism’ – it now takes longer for inputs to production to move to higher productivity firms, and it also takes longer for firms to catch up to the global frontier of performance and technology.[10] Evidence suggests that at least part of the decline in economic dynamism relates to declining competition in the economy.[11] Regulatory barriers also appear to have played a role in Australia, notably in the construction industry.[12] Other explanations include slowing human capital accumulation, declining trade integration, and mismeasurement.[13]

What does the recent subdued growth in productivity mean for our assessment of economic conditions? While productivity growth is associated with growth in incomes and wages over the longer run, in the short run there can be material divergences between these variables.[14] Over the past year or so, real average hourly earnings in the economy have grown faster than labour productivity. This exerts upward pressure on firms’ unit labour costs and is consistent with our assessment that labour market conditions are still tight, notwithstanding some easing in those conditions over the past couple of years.

What will happen from here? Our latest forecasts in the Statement on Monetary Policy incorporate a pick-up in productivity growth over the next couple of years, which would add to the economy’s supply capacity and help alleviate cost pressures. But there is considerable uncertainty around this projection. If productivity growth remains weak, the near-term outlook will depend critically on how the economy adjusts. If growth in demand is also weaker and wages adjust quickly to this slower growth in the supply capacity of the economy, there might not be a material impact on cost pressures. But if demand picks up as expected or wages adjust slowly to continued weak productivity outcomes, cost pressures could be higher than we expect. We will continue to monitor these developments carefully, alongside the full range of indicators we use to assess current economic conditions.

Concluding remarks

To conclude, productivity matters because it is a key driver of economic living standards. Over the longer run, higher productivity growth expands the supply capacity of the economy and supports growth in incomes, wages and aggregate demand. In the short run, however, there can be meaningful divergences in the growth rates of these important macroeconomic variables. Recent weak growth in productivity has constrained growth in aggregate supply. Whether productivity growth improves from here and how the economy adjusts are important questions for the economic outlook.

Thank you for your time today. I look forward to your questions.

Endnotes

I would like to thank Angelina Bruno, Jonathan Hambur and Kevin Lane for their help preparing this speech, and also Sarah Hunter, Natasha Cassidy, Leon Berkelmans, Tim Taylor, Thomas Williams, Oscar Douglas, Martin McCarthy, Michelle Bergmann, Stephanie Parsons, Gordana Peresin, Christopher Schwartz, David Bold, Hamish Sullivan, Georgia Face and Kieran MacGibbon for their comments and contributions. [*]

The measure of ‘wages’ used for these comparisons is average hourly earnings from the national accounts. It captures various forms of labour compensation (such as base salaries and wages, bonuses and overtime), as well as compositional change in the economy (such as workers moving from lower paying jobs to higher paying jobs). [1]

These averages exclude the farm sector, owing to challenges around measuring labour income in that sector. [2]

Productivity Commission (2024), ‘Higher Education and Productivity’, Speech by Danielle Wood, Chair, at the National Conference on University Governance, 23 October 2024, Canberra. [3]

In terms of the productivity cycles in Graph 5, the Australian Bureau of Statistics (ABS) determines productivity cycle peaks by comparing the annual MFP estimates with their corresponding long-term trend estimates. The peak deviation between these two series are the primary indicators of a growth cycle peak, although economic conditions at the time are also considered. [4]

See Bruno A, J Hambur and L Wang (2024), ‘Measuring Labour Quality in (Closer to) Real Time Using Emerging Microdata Sources’, Paper for Joint ABS-RBA Conference on Human Capital, June; Wiley G and L Wang (2024), ‘Skills Match Quality Following the COVID-19 Pandemic’, RBA Bulletin, July. [5]

These calculations include housing services as part of the market sector. [6]

See Luo Q (2020), ‘Hospital Output Measures in the Australian National Accounts: Experimental Estimates, 2004-05 to 2017-18’, ABS Paper, Canberra; Annabel J (2020), ‘Non-market Output Measures in the Australian National Accounts: Conceptual Framework Enhancements’, ABS Paper, Canberra. [7]

See Productivity Commission (2024), ‘Advances in Measuring Healthcare Productivity’, Research Paper, Canberra. [8]

The market sector excluding mining accounted for a little over two-thirds of total employment in 2024 and around two-thirds of gross value added in the Australian economy. See Australian Bureau of Statistics (2024), ‘Labour Account Australia, September 2024’, Canberra and Australian Bureau of Statistics (2024), ‘National Income, Expenditure and Product, September 2024’, Canberra. [9]

On reallocation, see Hambur J and D Andrews (2023), ‘Doing Less, with Less: Capital Misallocation, Investment and the Productivity Slowdown in Australia’, RBA Research Discussion Paper No 2023-03; Andrews D and D Hansell (2021), ‘Productivity-Enhancing Labour Reallocation in Australia’, Economic Record, 97(317), pp 157–169. On convergence, see Andrews D, C Criscuolo and P Gal (2016), ‘The Best versus the Rest: The Global Productivity Slowdown, Divergence across Firms and the Role of Public Policy’, OECD Productivity Working Paper No 5; Andrews D, J Hambur, D Hansell and A Wheeler (2022), ‘Reaching for the Stars: Australian Firms and the Global Productivity Frontier’, Treasury Working Paper No 2022-01. [10]

Hambur J (2023), ‘Product Market Competition and its Implications for the Australian Economy’, Economic Record, 99(324), pp 32-57. [11]

On the construction sector, see Productivity Commission (2025), ‘Productivity in the Construction Sector’, Inquiry Report No 105, Canberra. On occupational entry regulations, see Bowman J, J Hambur and M Markovski (2024), ‘Occupational Entry Regulations and Their Impact on the Australian Labour Market’, Treasury Working Paper No 2024-02. [12]

On slowing human capital accumulation, see OECD (2024), ‘Human Capital Accumulation and Productivity Growth’, OECD Economics Department Working Paper No 1602, pp 1–60. On trade, see Goldin I, P Koutroumpis, F Lafond and J Winkler (2024), ‘Why is Productivity Slowing Down?’, Journal of Economic Literature, 62(1), pp 196–268. Note that it has been argued that mismeasurement is unlikely to account for the slowdown globally and in Australia. See Syverson C (2017), ‘Challenges to Mismeasurement Explanations for the US Productivity Slowdown’, Journal of Economic Perspectives, 31(2), pp 165–186; Burnell J and M Elsnari (2020), ‘Mismeasurement and the Australian Productivity Slowdown’, Australian Economic Review, 53(4), pp 567–589. For a broader discussion of the slowdown and its causes, see Duretto M, O Majeed and J Hambur (2022), ‘Understanding the Productivity Slowdown: A Comprehensive Analysis’, Treasury Working Paper No 2022-05. [13]

A caveat to this shorter run analysis is that data on productivity and average hourly earnings can be volatile, which is why we tend to look at productivity over ‘cycles’ that span a number of years. So we need to be cautious when comparing growth rates over short periods. [14]