Watch video: Understanding the Journey to Full Employment

I would first like to pay respect to the traditional and original owners of this land, the Gadigal people of the Eora Nation, to pay respect to those who have passed before us and to acknowledge today’s custodians of this land. I also extend that respect to any First Nations people joining us here today.

Today I’d like to talk about the labour market. Employment is crucial to our living standards and our lives. A job is not just a source of income. For many of us it is also a key part of our identity and way of life. Those that experience a period of unemployment can suffer for a long time afterwards, and the economy can ‘lose’ their skills and expertise.

The Reserve Bank Board has historically set monetary policy to achieve both low and stable inflation and full employment. Our full employment mandate became more explicit in the updated Statement on the Conduct of Monetary Policy that was agreed between the Reserve Bank Board and the Treasurer late last year.^[1]

This morning I’d like to unpack our view of the labour market – how we’ve interpreted recent data, and what we expect over the next couple of years. But first I want to outline what we mean by and how we assess full employment. Then I’ll follow with our assessment of labour market dynamics – whether we think they’ve changed, and where we’ve been somewhat surprised recently. Finally, I’ll wrap up by talking about our outlook and the risks and uncertainties we’re closely monitoring.

What do we mean by ‘full employment’?

First, let’s start with our definition of full employment. Then I’ll unpack some of the complexities around how we measure and evaluate it over time.

We define full employment as the maximum level of employment that is consistent with low and stable inflation.^[2] When the economy is at full employment, demand and supply in the labour market are in balance. In turn, this leads to wages growing at a rate consistent with achieving the inflation target, after accounting for growth in productivity. In the medium term, this coincides with a balance between demand and supply in the markets for goods and services. And over time, the absolute level of full employment will typically rise, underpinned by growth in the working-age population.

Our full employment objective is easy to describe, but it is difficult to measure or capture in a single metric. So, it may not be a surprise that we use a range of methods and indicators to understand how the current state of the labour market compares to full employment, and to understand how much uncertainty there is around our assessment.^[3][4]

While the unemployment rate is a key metric in our analysis, we consider a range of different measures to holistically assess the labour market. These include leading indicators such as firms’ employment intentions, measures of unmet demand such as vacancies, and other cyclical measures of underutilisation such as the medium-term and youth unemployment rates, and the underemployment rate (Graph 1).

The RBA also uses a suite of models to estimate the unemployment rate and hours-based underutilisation rate (a broader measure of spare capacity which captures the shortfall in hours worked from unemployment and underemployment) that would be consistent with full employment. These models use wages growth and inflation to provide us with information on the balance between demand and supply in the labour market and in the economy. If inflationary pressures appear persistent, it may suggest that there remains excess demand in the economy and labour market. Deviations of the actual unemployment and hours-based underutilisation rates from these estimates – namely the unemployment and underutilisation gaps – provide yet another metric of the degree of tightness in the labour market.

I want to draw out another critical point. The level of full employment can change over time as the structure of the economy and labour market evolve. The maximum number of employed people that is consistent with the RBA achieving its price stability mandate will increase as the population grows or if the labour force participation rate rises (Graph 2). So, it is also useful to monitor trends in the employment-to-population ratio and the participation rate.

So what does all this mean for our current assessment of the labour market? Taking into account the methods and indicators that I’ve just taken you through, our current assessment is that the labour market is operating above full employment but has moved towards better balance since late 2022.

In the context of the first graph that I showed, we can see that a range of labour market indicators have eased since late 2022, at which time many of these indicators suggested conditions were very tight. Our models also suggest that the unemployment and underutilisation gaps have narrowed over this period, which also points to a labour market that is moving towards full employment. This assessment is based on what we know now. And from here, the return to full employment could play out in many ways. This will be determined by developments across the economy and in the labour market. So monitoring these developments is a crucial part of how we assess the overall outlook.

Underlying dynamics of the labour market

As many of you will know, when the economy begins to soften, the first thing firms generally do is to pull back on unmet demand for workers. That is, they recruit less intensely and put out fewer new job adverts, and cancel pre-existing vacancies. A softening of the labour market through this channel is not captured in data series such as the level of employment or the unemployment rate. Overall demand for labour can be falling even when employment is rising.

During the current cycle, we first observed declines in both job vacancies and measures of employment intentions from business surveys and the RBA’s liaison program in 2022. At the time, conditions were very tight and there were many vacant positions. As momentum in demand started to moderate, firms responded to the slowing in momentum by withdrawing positions or posting fewer vacancies. The vacancy rate has subsequently fallen since mid-2022 but remains above its pre-pandemic level.

To date, the decline in vacancies has coincided with only a modest increase in the unemployment rate. The typically inverse relationship between vacancies and unemployment is what economists call the ‘Beveridge curve’. The economy appears to have been operating on the steeper portion of the curve of late, where a fall in vacancies can occur without a large increase in unemployment (Graph 3). We also can’t see clear evidence that the curve permanently shifted over the pandemic years. For example, if the curve had shifted out, this would suggest that matching workers to jobs had become harder, which could be because firms are demanding new or different skills, or because workers weren’t able to move to the sectors or regions where there are job openings.

Our current assessment, however, is that this hasn’t happened. In other words, we think the labour market today is operating broadly the way it did prior to the pandemic.

While we have not seen a sharp lift in the unemployment rate, the decline in demand for labour and firms’ hiring intentions has translated into a lower rate of new hiring and slower employment growth. We can see this in the rate at which jobseekers have been able to find work. The job-finding rate rose to multi-decade highs in 2022 before falling back, though it remains well above its 2010s average (Graph 4). This observation is consistent with people taking a bit longer to find work if they are looking for it. All else equal, this dynamic contributes to a higher unemployment rate.

Firms also respond to slowing growth in demand by reducing the hours of their existing workforce. We know this is easier for some roles and sectors than others. We may therefore observe a reduction in average hours worked per employee.

Average hours worked did decline over much of 2023. But the level of average hours worked has not fallen materially since late last year and looking through the volatility we’ve seen this year, has remained broadly steady since (Graph 5). The data also highlight the volatility we’ve seen in average hours worked over the past four years or so, and understanding how it might evolve from here is a key uncertainty in our forecasts.^[5]

Similar to average hours, other measures of labour underutilisation have also not increased noticeably in recent months. As an example, the underemployment rate, which captures workers who would like to work more hours, rose in 2023 but has been relatively steady since earlier in the year.

The limited signs of further easing in the latest observations for variables such as average hours worked and the underemployment rate at a time when economic growth has continued to slow is somewhat surprising and is a key uncertainty in the outlook. However, taken together, they suggest that conditions in the labour market remain tighter than implied by full employment.

In a downturn, particularly in a more severe downturn, firms may also turn to layoffs to reduce labour costs. This would further raise the unemployment rate. Past downturns tell us that firms resist laying off staff, if possible, as they try to avoid the costs associated with rehiring and reskilling workers as the economic outlook improves.

The layoffs rate has been trending up recently, but it remains very low by historical standards (Graph 6). Our assessment is that this is consistent with the labour market loosening from very tight conditions, but it is still operating above full employment. Changes to the layoffs rate is another key metric we monitor.

Conditions in the labour market are also typically reflected in the pace of wages growth, and tightness in the labour market has contributed to strength in wages growth in recent years. More recently, there are some signs that the easing in the labour market since late 2022 has started to flow through to wage outcomes for workers. Nominal wages growth appears to have passed its peak and has moderated since late last year, though wages growth remains solid (Graph 7). Having said that, real wages, which matter more for households, have been weaker, reinforcing the costs of high inflation.^[6] Moreover, growth of labour costs, as measured by growth of labour costs per unit of output, has also declined (in part due to an improvement in productivity outcomes).

The recent easing in historical context

It’s important to put the recent easing into the context of historical episodes, so we can better forecast how the labour market is likely to evolve from here. Overall, our current assessment is that the recent easing in labour market conditions has, to date, been similar to mild downturns in Australian history (Graph 8).

Excluding the pandemic period, downturns in economic activity and their associated impacts on labour market outcomes were larger in the early 1980s and 1990s compared to downturns in the 2000s and 2010s. This is shown by the magnitude of declines in total hours worked which accounts for both changes in employment and average hours worked.

In the current episode, the decline in total hours worked has come from reductions in average hours worked rather than declines in employment, repeating the pattern seen in the most recent episodes. A likely explanation for this is that more recent downturns have been smaller in magnitude, but it may also partly reflect increased flexibility in the labour market. This has allowed firms to contain labour costs while retaining employees, and avoid the costs associated with layoffs and rehiring.^[7] It has also meant fewer layoffs and smaller increases in the unemployment rate, which has kept more workers in jobs.

With the recent adjustment in total hours arising via average hours worked, total employment has continued to increase, notwithstanding the unemployment rate rising from late 2022 onwards. Aggregate employment growth has been supported by the rebound in migration since late 2021, which has contributed to both an increase in the supply and demand for labour.

Employment growth has also been uneven across sectors, with aggregate employment supported by strong growth in the health, education and public administration sectors. Importantly, the labour market appears to be doing a good job of facilitating this flow of workers to these industries.^[8] Our preliminary analysis of the healthcare sector, for example, suggests that it has drawn significantly from those who were outside the labour force or unemployed, while also bringing in some people previously employed in other industries.

What has been surprising?

I don’t want to create an impression that we know everything about the labour market, particularly about how it will evolve from here.

It’s true that recent unemployment rate outcomes have not been unusual given the historical relationship between changes in the unemployment rate and GDP growth rates (Graph 9).

What has been more surprising is the strength in the labour force participation rate and, as discussed earlier, the limited easing more recently in some measures such as the underemployment rate and average hours worked. This is occurring against a backdrop of slow momentum in the economy, a decline in vacancies and a gradually rising unemployment rate.

The participation rate has continued to trend upwards, reaching a record high in July. This has not been the experience in most other peer economies (Graph 10). Whereas participation rates amongst this group have largely begun to stabilise, in Australia the participation rate has continued to increase steadily in the post-pandemic period.

While the slowing in the economy may have weighed on the participation rate (all other things equal), other factors have likely provided support, including the longer-run trend towards greater female participation. We have also seen the share of employed people with multiple jobs continue to trend upwards since the pandemic, notwithstanding the decline in the latest quarterly data (Graph 11).

So where to from here?

To sum up, conditions in the labour market have eased since late 2022, but our assessment is that the labour market is still tight relative to full employment.

We expect the demand for labour to grow at a slower pace relative to the supply of labour in the coming quarters, gradually bringing the labour market into better balance. Our view is that some of this slowing in labour demand is likely to occur via a decline in average hours (Graph 12).

We also expect employment to continue to increase, but at a slower pace than population growth. In this view of the outlook, measures of underutilisation – including the unemployment rate – are expected to continue rising gradually from here, before stabilising as the pace of growth in GDP picks up to be broadly consistent with the economy’s underlying trend pace of growth.

As my colleague Andrew Hauser recently pointed out, the outlook is highly uncertain, and we should be humble about our ability to predict the future. If we can be confident about anything, it’s that our forecasts will be wrong at least in some way. We therefore spend a lot of time thinking about how and why we could be wrong, including by considering scenarios where the economy evolves differently to our base case forecast and by continuously updating our assessment of risks as economic conditions change. This is particularly important given that changes in economic conditions can take time to flow through to changes in labour market conditions.

It’s possible that employment growth slows, and the unemployment rate rises more quickly than we expect. As previously noted, various leading indicators, such as vacancies, are continuing to ease, suggesting further softening in the labour market from here. Our view is that further falls in vacancies can still occur alongside a relatively modest increase in the unemployment rate. Compared to other economies that have seen larger increases in their unemployment rates, the Australian economy has a higher ratio of vacancies to unemployment, relative to historical experience, which suggests that there is space for vacancies to fall further without a sharp increase in the unemployment rate (Graph 13).

This outcome is consistent with our assessment that we are still on the steeper part of the Beveridge curve that I showed earlier. But the slope of the Beveridge curve is highly uncertain, and our view could easily be wrong.

Our central forecast is also predicated on the assessment that few firms are operating with excess labour – engaging in so-called ‘labour hoarding’. If it turns out that more firms are currently hoarding labour, then we may expect to see a larger pick-up in the unemployment rate as firms reduce headcount to cut costs given the backdrop of weak growth in demand. The implications of this alternative were explored in a scenario in the RBA’s August Statement on Monetary Policy.

It is also possible that our assessment is wrong in the other direction. Conditions may be tighter than we expect, or demand for labour could grow more strongly than we anticipate.

Forecasting the labour market is a difficult but important challenge. Our current assessment is that labour market dynamics haven’t fundamentally changed, although we have been surprised by some of the recent data. While our forecasts will almost certainly be wrong in some way, we remain humble about our ability to predict the future, and we continue to update our views based on the incoming data and our ongoing analysis.

Endnotes

I would like to thank Joyce Tan, Jessica Marnie, Kevin Lane, Anirudh Yadav, Michelle Bergmann, Benjamin Beckers, Kieran MacGibbon, Andrew Hauser, Chris Kent, Brad Jones and Michele Bullock for their contributions and comments on the text. [*]

The Treasurer and the Reserve Bank Board (2023), Statement on the Conduct of Monetary Policy, December. [1]

Bullock M (2023), ‘Monetary Policy in Australia: Complementarities and Trade-offs’, Speech at the Commonwealth Bank Global Markets Conference, Sydney, 24 October. [2]

Ballantyne A, A Sharma and T Taylor (2024), ‘Assessing Full Employment in Australia’, RBA Bulletin, April. [3]

There is no mechanical link between these indicators and our overall assessment of labour market conditions relative to full employment. For example, these variables may have trended higher or lower over time, meaning that we should not expect the current outcome to return to the middle of its historical range. [4]

The recent volatility notwithstanding, average hours worked per employed person in Australia has declined over many decades, consistent with the rise in part-time work and the increased participation of women in the labour force. [5]

For more details on the cost of inflation, see Bullock M (2024), ‘The Costs of High Inflation’, Keynote Address to the Anika Foundation Fundraising Lunch, Sydney, 5 September. [6]

Bishop J, Gustafsson L and Plumb M (2016), ‘Jobs or Hours? Cyclical Labour Market Adjustment in Australia’, RBA Research Discussion Paper No 2016-06. [7]

Analysis by the Commonwealth Treasury suggests that there have been improvements in how well the labour market is able to match unemployed individuals to available jobs (‘matching efficiency’) in recent years. Matching efficiency was only slightly below its long-run average in 2023. See Mackey W (2024), ‘Labour Market Matching Across Skills and Regions in Australia’, Treasury Round Up, March. [8]