Speech Adjustments in the Labour Market
Assistant Governor (Economic)
Address to the Economic Society of Australia (Qld) Luncheon
Let me start by thanking the Queensland branch of the Economic Society of Australia for hosting this event today, and special thanks to Michael Knox and Tim Wonhof for their assistance.
The Australian economy has been growing at a moderate pace since commodity prices peaked around four years ago. While mining investment has declined substantially since then and has further to fall, resource exports have grown strongly and are expected to do so for some time. Fiscal consolidation at state and federal levels has weighed somewhat on growth. And the exchange rate has been relatively high compared to what one might have expected given these circumstances. At the same time, very low interest rates have been supporting demand.
So there have been a range of different forces affecting the economy. Overall, spare capacity in product and labour markets has gradually increased. The unemployment rate rose from around 5 per cent in 2011 to about 6¼ per cent towards the end of last year (Graph 1). Also, the participation rate and the ratio of employment to the working age population had fallen to relatively low levels. But in the face of significant structural change in the economy, the labour market has been adjusting more smoothly over the past year than we had been expecting.
When we put our forecasts together earlier this year, we expected moderate growth of economic activity for some time. The GDP growth that has been recorded suggests that we have been about right on this so far. We also thought this would lead to a further gradual increase in the unemployment rate and be accompanied by subdued growth of employment.
But, since then, the labour force data have been a bit better than expected. While the unemployment rate remains high, looking through the month-to-month volatility, it has been little changed since the latter part of last year. In addition, employment growth has picked up and has been running ahead of the growth in the working-age population.
How can we reconcile better-than-expected labour market outcomes with only moderate growth in GDP? The focus of my remarks today is on four possible explanations for this:
- First, recently released data revealed that population growth, and hence the growth of the labour force, has been a bit lower than we'd assumed. In the face of fewer employment opportunities over recent years, particularly in resource-related industries, there has been a decline in the (net) number of immigrants arriving to seek work. While we'd expect the labour supply to respond to economic conditions in Australia relative to those overseas, the adjustment has been more pronounced than earlier anticipated.
- Second, wage growth declined as spare capacity in the labour market increased. This is also unsurprising, but the adjustment has been larger than suggested by historical experience. This has probably encouraged more employment growth than otherwise and improved the cost competitiveness of Australian labour. It also acts as a signal to encourage adjustment in the supply of labour.
- Third, the composition of growth may be playing a role in boosting demand for labour. While aggregate economic activity has grown at a below average pace, conditions in labour-intensive parts of the economy seem to have improved.
- Finally, the key sources of data – the labour force survey and the national accounts – are both subject to measurement error and are often revised. It is possible that the current data are providing a noisy signal of what is actually happening.
I'll go over each of these points in turn and then come back to a discussion of the likely implications of these developments for the economic outlook.
Changes in Population Growth
As recently as May, official data suggested that the working-age population had been growing by about 1.7 per cent per annum, and it was expected to continue at about that rate in the foreseeable future. However, the most recent data from the ABS suggest that total population growth had dropped quite noticeably over the past year or so, from 1.8 per cent over 2012 to 1.4 per cent over 2014 (Graph 2). Estimates of the working-age population are expected to be revised down accordingly in coming months.
The decline in population growth was mainly the result of a decline in net immigration. This appears to be related to the weakening in Australia's labour market conditions relative to those of other countries. Of course, people on temporary skilled-work visas (so-called ‘457s’) leave for home when their jobs end. But there has also been a large decline in net immigration from New Zealand. Labour market conditions in New Zealand tightened at the same time that spare capacity in the Australian labour market increased. The reconstruction activity in Christchurch took off around the time that construction in Australia's resource sector was winding down. Indeed, the key mining states of Queensland and Western Australia have seen the largest declines in net immigration. It also appears that Australia's intake of international students has not picked up to the extent previously anticipated.
The unexpected slowdown in population growth implies somewhat less rapid growth of our labour force than otherwise. This means that the GDP growth that we have recorded may have been closer to the recent growth in the economy's productive capacity than previously thought. If so, that would have left the economy with a little less spare capacity (a lower unemployment rate) than had been expected. Yet this can only account for part of the unexpected adjustment in the labour market. Most notably, labour demand has also picked up. That is evident in the rise in employment relative to the population.
The Role of Wage Growth
One possible explanation for the pick-up in labour demand relates to the large decline in wage growth over recent years (Graph 3). The behaviour of wages during the current episode has been comparable to the experience around the 1990s recession. This is true of nominal wage growth and growth in the cost of the labour required to produce a unit of output – so-called unit labour costs. The decline in real wage growth has also been of a similar magnitude to the early 1990s.
While we would normally expect wage growth to decline after a period of subdued labour demand, the decline over recent years has been larger than suggested by historical experience. In particular, although the unemployment rate has increased by much less than during the early 1990s recession, the decline in wage growth during the two episodes has been similar (Graph 4).
There are a number of reasons why wage growth may have been so responsive this time around. For example:
- During the boom in commodity prices and mining investment, wage growth increased noticeably and the exchange rate appreciated significantly. This left our wage costs in foreign currency terms quite high once commodity prices had peaked. The subsequent fall in commodity prices was associated with relatively low growth of wages over recent years which, alongside some depreciation of the exchange rate, is helping to restore earlier levels of competitiveness.
- Wages may have become more flexible over time. It may be that there has been some general shift in the bargaining power of labour. Also, those industries most exposed to the declines in commodity prices and mining investment have a high share of individual employment contracts. Furthermore, the relatively long span of the current episode means that most employment contracts have been renegotiated during the period of subpar economic growth over recent years.
Whatever the reason for the very low growth in labour costs, wage flexibility has assisted with the labour market adjustment. Changes in relative wages act as an incentive for labour to move. Strong increases in wages in the resource sector (in both domestic and foreign currency terms) helped to attract the labour needed to undertake the substantial increase in mining investment (including from overseas). All of these things are now adjusting in the other direction. The weaker wage growth and decline in job opportunities relative to many other countries has probably contributed to the lower growth in the population and labour supply. In addition, low wage growth across the economy has enabled firms to employ more labour than would otherwise have been the case. Even so, my sense is that low wage growth only goes some way to explaining the recent pick-up in labour demand.
While aggregate GDP growth has been moderate, the composition of economic activity may have underpinned a rise in the demand for labour of late. Growth in consumption, dwelling investment and net service exports have increased over the past year even though GDP growth has eased back a touch in year-ended terms (Graph 5).
Surveys of businesses suggest that for firms providing services to households, conditions increased substantially from mid 2013 and have been well above average over the past couple of years (Graph 6). Surveyed conditions are also above average for those firms providing business services. Meanwhile, conditions for firms producing or distributing goods remain a bit below average. The surveys of business conditions by sector line up reasonably well with employment growth across sectors.
As an interesting aside, variation in business conditions and employment across sectors is consistent with the weakness in business investment. The service sectors, on average, have relatively low capital-to-labour ratios compared to those producing or distributing goods (Graph 7). So in order to satisfy increased demand for services, firms in those sectors have hired more workers but they didn't need to invest much in machinery & equipment or buildings & structures (compared with firms producing goods).
This compositional change in demand towards more labour-intensive industries is part of a longer-running trend. But there is also a cyclical element to it as mining investment unwinds and growth of dwelling investment and consumption (which is increasingly dominated by services) has picked up. Also, the decline in the exchange rate has encouraged Australians and foreigners to direct more of their spending to Australian tourism, education and business services.
It's possible that the decline in population growth, very low wage growth and compositional change can explain the adjustments evident in the labour market data even though growth of aggregate economic activity has remained moderate. It is also possible, however, that the picture before us of recent labour market outcomes and economic conditions more generally are being clouded by usual measurement difficulties.
There is naturally a degree of statistical noise in the labour force survey, as with any survey data, so it might be providing a noisy read of true labour market conditions. While it is possible that the labour force survey is overstating the true state of demand for workers, it is worth noting that the separate ABS survey of businesses' vacancies is broadly consistent with the improvement in labour market demand of late. Vacancies have picked up for the nation as a whole since 2013, and most of the growth has been driven by New South Wales and Victoria, and rising vacancies in the service sector (Graph 8). Meanwhile, vacancies in Queensland and Western Australia, and for the goods sector, have been weak.
Similarly, the national accounts data might be understating the growth of economic activity over the past year or so. Some research by my colleagues shows how sizeable these revisions can be. For example, since 1998, the first estimate of year-ended GDP growth has been revised, on average, by ½ percentage point over subsequent years.
While we can't be sure of the recent past, more difficult still is to know what the future holds.
The Economic Outlook
We have recently published an account of our latest economic forecasts in the August Statement on Monetary Policy. I won't repeat the details of those forecasts, but I want to highlight the implications of the four issues I've just discussed.
Changes in Population Growth
The implications of lower population growth for the outlook for spare capacity depends on its effect on aggregate demand and aggregate supply.
Lower population growth implies slightly less growth in aggregate demand for goods and services. At the margin, there will not be as many new residents spending on items for consumption or needing housing. Nor will there be as many new workers that we'd need to equip with capital to do their jobs.
Despite the downward revisions to a number of elements of aggregate demand, the unemployment rate is forecast to be a little lower than previously anticipated. In part, this owes to the slightly better starting point for the unemployment rate than earlier assumed. Also the unemployment rate is forecast to remain little changed from levels of recent months over the course of the rest of this year and next. This is because the change in the growth of aggregate demand is expected to be broadly matched by lower growth of the economy's productive capacity, which in turn follows from less growth in the labour force.
While that is what we expect, there is considerable uncertainty about both the rate of population growth and its effects on spare capacity. An example might be helpful here. Lower immigration may reflect a fall in the number of people coming to Australia who have a relatively high propensity to participate in the labour force, such as New Zealanders of prime working age. In that case, the implications for the economy's productive capacity will be larger than if lower immigration was mostly due to fewer international students. It is not clear whether these two groups would have the same effect on demand: both need to feed, clothe and house themselves, but both have different incomes and different propensities to spend and save.
Another source of uncertainty affecting the outlook for supply is productivity growth. That is especially hard to forecast at any time. What signal can we take from recent developments? Not much, is my sense. We know that the lower-than-expected population growth will (by itself) lead to a downward revision to employment growth. For the given growth in GDP that has been recorded of late, this implies that existing estimates of recent productivity growth are somewhat understated. In other words, they are likely to be revised a bit higher when the ABS revise employment growth. However, recent measures of productivity growth tend to provide a very noisy signal of future productivity growth. So we haven't changed our view on the outlook for productivity growth. We assume that it will be around its historical average.
The Role of Wage Growth
What about relevance of the low growth of wages for the economic outlook?
It is reasonable to be concerned that low wage growth might weigh on labour income thereby weakening the outlook for aggregate demand. But that is an incomplete description of a more complicated process. In particular, wages tend to lag the business cycle, not lead it. This makes sense for at least two reasons. First, many wages are determined by fixed-term agreements. So they take time to adjust to changing conditions in the labour market. Second, even after growth of output picks up, it can take time to absorb spare capacity in the labour market, and hence for wage pressures to build.
Given this, the low wage growth of recent years is perhaps best viewed as a response to the spare capacity that has built up in the labour market over time. If wage growth had not been so responsive, it is likely that employment would have been less and the unemployment rate higher than we've seen. So even though low wage growth works to constrain the growth of incomes for those who are employed, it also supports incomes by encouraging more employment than otherwise.
The outlook for the labour market will also depend on the composition of future labour demand. When looking for signs of a turnaround in economic activity, attention is often focused on the tangible things that tend to be measured more easily – the goods we produce and consume, and the machinery, buildings and infrastructure in which we invest. Yet a large and increasing share of our economic activity consists of the services we provide, use and enjoy as well as the investment in intangible capital, including human capital.
About half of us are now employed in the service industries. This has increased from about 45 per cent two decades ago. But over the past three years, the additional workers employed in service industries have outnumbered the extra workers employed in the goods sector by a factor of five to one (347 thousand compared to only 73 thousand). Those employed in services require relatively less capital to work with than workers employed in the goods sector. And as I mentioned, while some of the recent improvement in the services sector is likely to be part of a longer-running trend, part of it may also be cyclical. Signals from the various business surveys, the ABS labour force and vacancies data, can provide a timely guide as to the health of the increasingly important services sector.
This leads me back to the issue of measurement. Output and productivity in the services sector is more difficult to measure than it is for the goods sector. But with services becoming an increasingly large share of the economy, finding better ways to measure the contribution of services to the economy in real terms is going to become increasingly important in understanding our performance. The key issue is that improvements in the quality of services are hard to measure, but these are an important source of improvements in our economic wellbeing.
In summary, the economy has been adjusting to the various headwinds it has faced in recent years. Even though measured GDP growth has remained moderate over the past year or so, the unemployment rate appears to have stabilised and demand for labour has picked up.
Flexibility in labour markets has played a role in these adjustments. One aspect of that has been some change in the growth of the labour supply via changes in the extent of immigration. Another aspect of the adjustment has been the very low growth of wages, which has helped to support a pick-up in employment growth.
The improvement in labour demand is consistent with the rise in consumption growth, strong dwelling investment and an increasing contribution to GDP from net services exports. It is also apparent in improvements in survey measures of business conditions, vacancies and employment in the service sectors.
While there is uncertainty around our forecasts, including because of the usual noise in data, these developments suggest that there will be slightly less spare capacity in the labour market than earlier anticipated. In particular, we expect the unemployment rate to be little changed from recent levels over the next 18 months or so, before declining in 2017.
I thank Kathryn Davis, David Jacobs, Alexandra Rush, Martin McCarthy, Andrew van der Saag and Madeleine McCowage for excellent assistance in preparing these remarks. [*]
Net immigration had been very high for a number of years and was an important means of addressing skill shortages. Immigration was especially significant in Queensland and Western Australia, where demand for labour from the resource investment boom was most concentrated. As the mining investment boom has wound down, and demand for labour has weakened, these states have also seen the largest declines in net immigration. For more detail see RBA (2015), ‘Box D: Implications of Lower Population Growth for the Australian Economy’, Statement on Monetary Policy, August, pp 44–46. 
This discussion is based on data from Statistics New Zealand and the Department of Education and Training. More detail on net immigration by visa category will be released by the Department of Immigration and Border Protection in due course. 
See Jacobs and Rush (2015) for a more comprehensive discussion of these points. 
This sort of adjustment was to be expected. For a conceptual framework see Plumb M, C Kent and J Bishop (2013), ‘Implications for the Australian Economy from Strong Growth in Asia’, RBA Research Discussion Paper No 2013–03. 
See Kent C (2014), ‘Ageing and Australia's Economic Outlook’, Address to the Leading Age Services Australia (LASA) National Congress, Adelaide, 20 October. 
The labour force series do not yet reflect recently released data showing lower rates of population growth, which implies that the number of people in each labour force category and their growth rates are overstated. However, measures expressed relative to the size of the labour force or population (such as the unemployment rate or employment-to-population ratio) are likely to be largely unaffected by revisions that will flow from the new population data. 
While the raw numbers of employees have not yet been revised in light of new population estimates (see footnote 7), the relative status of each sector is not likely to change significantly.