The Long Cycle in Australian Construction Productivity

 

Over the last 15 years, Construction productivity in Australia has gone through a long cycle and ended up more or less where it began. The Australian Bureau of Statistics Construction labour productivity index rose from 92.88 in 2006-07 to 95.36 in 2010-111, before rapidly increasing to 115.94 in 2013-14, then falling to 99.1 in 2021-22. What explains the 10 year cycle between 2011 and 2021?

Productivity estimates require both a measure of labour inputs, such as hours worked or people employed, and a measure of output, called Industry value added (IVA, the difference between total revenue and total costs). IVA is then adjusted for changes in prices of materials and labour to estimate Gross value added (GVA) using price indexes. The ABS has a Construction industry labour productivity index, but does not have separate indexes for the different construction industry sectors.

However, using GVA data, estimates of productivity for the industry sectors of Engineering construction, Building construction and Construction services can be found. The GVA data comes from the ABS National Accounts (chain volume measures of economic activity). The number of people employed includes all workers in June each year, and comes from ABS Australian Industry. The construction work done data is from the ABS chain volume Value of Construction Work Done, which is expenditure on construction adjusted for inflation.

GVA per person employed is a useful proxy for industry productivity. As a combination of real output and employment, GVA per person employed looks like a measure of productivity and, while not precise, it is indicative of industry trends. As a measure of productivity, annual GVA per person employed follows a similar path to the ABS productivity index, which uses the number of hours worked for labour input. If output is increasing faster than employment, labour productivity will also increase.

In Figure 1 Construction GVA per person employed in thousands of dollars per person is compared to the ABS labour productivity index for Construction from their Estimates of Industry Multifactor Productivity. There is a good match between the two because they both use industry GVA for output and are both based on the 2020-21 year, although the long cycle between 2011 and 2021 is more pronounced in the GVA per person data. The ABS labour productivity index is 2020-21 = 100, and went from 92.88 in 2006-07 to 115.94 in 2013-14 before falling to 99.1 in 2021-22.

Figure 1. Construction Gross Value Added per Person Employed and Labour Productivity

In Figure 2 the Value of construction work done is shown with GVA per person employed, and there is a clear relationship between the two. Over the 15 year period Construction work done peaked twice, first in 2014 during the mining boom and then in 2018 with new infrastructure projects. Construction GVA per person followed a similar trend to the rises and falls in work done. 

Figure 2. Gross Value Added per Person Employed and Construction Work Done

Significantly, Figure 2 shows GVA per person employed also peaked in 2014 as it followed changes in work done. Productivity increased when construction was rising after 2011 during the mining boom, and then decreased as the mining boom ended. The first stage of this long-run cycle of increasing construction work done and GVA per person employed in construction was entirely due to the large number of large, capital intensive resource projects completed during the mining boom between 2011 and 2017, followed by transport and energy sectors commencing in the second stage from 2018.

Engineering Construction

The mining boom started in the 2000s but took off after 2010 with many large energy and resource projects. As Figure 3 shows, the value of Engineering work done doubled between 2007 and 2013, and as it did Engineering GVA per person employed increased by around 20%. Engineering GVA per person followed the rise and fall and subsequent fall in the value of work done, and is now well below the peak years. The value of Engineering work done includes associated expenditure on machinery and equipment, so a lot of the increase was due to the capital investment required for the LNG plants, offshore platforms, railways and new mines constructed.

Figure 3. Gross Value Added per Person Employed and Engineering Work Done

Building Construction and Construction Services

Using Australian Industry employment data, GVA per person employed can also be found for both Building construction and Construction services as well as Engineering. Here a slight decline in Building has been offset by a small rise in Construction services GVA per person over the 15 years, as GVA per person since 2006-07 in Building has declined slightly, from $179,000 to $148,000 in 2021-22, while it increased for Construction services, from $84,000 to $102,000. As shown in Figure 4, the level of GVA per person in Construction services is around two-thirds of the level for Building construction, because Construction services are generally labour intensive and will therefore have a lower value of output per person. 

The range in output per person employed between Engineering, Building and Construction services reflects the differences in capital requirements of these three sectors, and expenditure on purchases of software, equipment and machinery by firms in the three sectors. The higher the capital requirements, or capital intensity, of an industry, the higher the level of output per person employed is expected to be, because workers with more capital are more productive. Both excavators and shovels require one operator, but the former shifts more soil.

Figure 4. GVA per Person Employed in Engineering, Building and Construction Services

The strong relationship between the value of Engineering work done and GVA per person seen in Engineering is not found for Building and the value of Building work done. Despite the increase in the value of Building work done after 2014 there was no increase in GVA per person, rather there was a slight decline as residential building increased during the transition after the mining boom, as shown in Figure 5. This may also be an indicator of increased use of prefabrication and offsite manufacturing reducing the value added of onsite work, but we have no reliable data on that. 

Figure 5. Gross Value Added per Person Employed and Building Work Done

For Construction Services there is real GVA but not real work done data, so total construction work done is used in Figure 6, as the trades work across Building and Engineering. Here, GVA per person does increase as work increases, but peaks in 2016 and then falls away. The number of people employed increased from 723,000 in 2015-16 to 854,000 in 2020-21, and may have contributed to the fall in GVA per person if that increase included many new workers with little experience and limited skills. There was a particularly big increase in employment in 2020-21, when over 60,000 more people were employed in Construction services, and employment grew much faster than work done. As a result, GVA per person dropped by $8,000 before recovering in 2021-22.

Figure 6. Gross Value Added per Person Employed and Total Construction Work Done

Conclusion

Over the last 15 years, construction productivity in Australia has gone through a long cycle and ended up more or less where it began. Rising from $115,000 per person employed in 2007 to $149,000 in 2014, an increase of 29%, GVA per person employed then started falling and was back to $125,000 per person in 2022.

This long cycle in Australian construction productivity followed changes in the value of Engineering work done. Productivity increased as Engineering construction rose after 2011, and then decreased as the mining boom ended. This long-run cycle in construction productivity was entirely due to the capital intensive resource projects completed between 2011 and 2017 by Engineering construction, where the machinery and equipment required for LNG plants, offshore platforms and new mines is included in the value of work done. This greatly increased the value of work done, but much of the increase was therefore not invested directly in construction work.

The strong relationship between changes in the value of work done and GVA per person seen in Engineering is not found for Building construction or Construction services. The higher capital intensity of Engineering work appears to lift GVA per person with increasing work done, as more expenditure on machinery and equipment is included in this sector compared to building.

Despite the increase in the value of building work done after 2014 there was no increase in Building GVA per person, rather there was a slight decline. For Construction services, GVA per person did increase to 2016 as construction work done increased, before falling away as the number of people employed increased faster than the value of work done.

What can be taken from this episode? Firstly, over 15 years there has been little change in overall construction labour productivity and, unlike Engineering, for Building and Construction services increasing work done has seen productivity fall slightly since 2014 as employment increased faster than output. Second, the lack of any real trend in construction productivity, despite changes in output and a continual increase in the number of people employed, suggests the industry is at the limits of efficiency, based on current technology. As output increases the number of people also increases, often a bit more than output but sometimes slightly less, so GVA per person is not improving. Therefore the industry may be somewhere close to the efficiency frontier in delivering projects, but there is no trend at either the industry or sector level of increasing productivity.

Australian Manufacturing of Prefabricated Buildings and Construction Products

 The extent of prefabrication used in Australian construction is unknown and unknowable

 

 

The Australian construction industry is supplied by an extensive manufacturing base that includes a wide and varied range of industries, producing machinery and equipment as well as materials like bricks, glass, concrete, steel and wood. In 2021-22 there were 133,216 people employed in construction related manufacturing in Australia.

 

The Australian Bureau of Statistics also includes prefabricated buildings in manufacturing, however, the data is limited to the relatively small number of firms that classify themselves as prefab manufacturers, and misses offsite work by firms that may be classified as building or trade contractors, architectural or engineering practices, or work done inhouse in other industries like tourism and aged care.  

 

Therefore, the actual extent and depth of prefabrication used in Australian construction is unknown, and with the data available is largely unknowable. With offsite manufacturing in general, and prefabrication in particular, seen as important to addressing the industry challenges of sustainability, productivity and skills, the lack of data on how many and what type of prefabricated buildings and components are produced each year in Australia is a significant gap in knowledge and understanding of the industry.

 

This research starts with the ABS Manufacturing industry data. It then looks at their estimates of the number of dwellings on manufactured home estates and the effects of misclassification on those estimates. The next question addressed is the number and type of firms producing prefabricated buildings. 

 

 

Construction Related Manufacturing

 

Australian Manufacturing is divided by the ABS into 19 subdivisions, with the subdivisions made up of groups of firms classified by similarities in their products or processes into classes. The ABS gives an industry class an ANZSIC four digit number, and that level of detail allows an estimate of the employment and output of construction related manufacturing to be made. The data comes from the ABS annual publication Australian Industry.

 

The largest of the relevant classes are for widely used materials like wood, concrete, steel and glass. Other classes include manufactured products like plaster and ventilation systems. Table 1 shows the industry classes identified as directly contributing to new construction, with the number of people employed in June, and in Table 2 their output as Industry value added (IVA) in current dollars. Figure 1 shows the totals.

 

Figure 1. Total employment and output

 

In 2021-22 construction related manufacturing was 16% of total manufacturing IVA and employment, having increased from 15% in 2014-15.

 

Firms self-select the ANZSIC industry code used to classify them into an industry class. Some building companies like Sekisui Australia, Hickory and Hutchinson have offsite facilities but are not manufacturers. A smaller builder that does some modular or offsite construction might be classed as a building contractor or a construction trade such as Carpentry services. Prefabricated buildings produced inhouse by an organisation in an industry like transport, tourism or retirement villages will not be included in manufacturing. 

 

Therefore, there is some give and take as regards to what is included in and excluded from these industries, an outcome of the ANZSIC classification system. Some building products are not included, like floorboards, carpets and insulation, because they belong to larger product groups and can’t be separated in the data. On the other hand, industries included here like Glass products and Paint and coatings supply a range of other industries besides construction.  

 

The four largest manufacturing industries supplying construction add up to nearly 75,000 people employed producing over $8 billion in added value in 2021-22, in Table 3. 

 

Prefabricated Buildings and Concrete

 

There are two industries producing prefabricated buildings, of wood and metal respectively. These have grown significantly since 2014, although the big jump in IVA in 2022 may be revised in next year’s release. Nearly 10,000 people produce wood and steel prefabricated buildings. The ABS does not have data on what types of buildings are produced (i.e. residential, commercial, institutional etc.). Wooden building prefab is a very small industry, in 2021-22 total income was only $688 million and IVA was $251mn, compared to prefab steel building with income of $3.5 billion and an IVA over $1.1bn. 

 

Concrete product manufacturing employed 7,670 people, a substantial industry that produces pots and bricks, but also prefabricated elements and buildings. The precast concrete industry is highly concentrated, with six major firms (ADBRI, Brickworks, CSR, CTC Precast, James Hardie, Holcim) and a large number of small and medium size firms around the country. 

 

The combined total of offsite construction in 2021-22 was over 17,000 people employed and an IVA of $2.4 billion. 

 

 

The share of wood and steel prefabricated buildings in total construction related manufacturing (in Figure 2) increased by nearly half between 2014-15 and 2021-22. This may be the strongest signal in this data of the uptake of modern methods of construction and increasing use of offsite construction. 

 

Figure 2. Wood and steel building’s share of total construction related manufacturing

 

 

How Many Manufactured Homes?

 

The 2021 ABS Housing Census dwelling location data includes manufactured home estates and long term residents in caravan parks, and there were over 10,000 of these manufactured houses in Australia, and another 2,000 townhouses and apartments (Table 5). Unfortunately, the 2016 Census housing data did not include this category. The ABS had a separate category for Retirement villages in the 2021 Census, with over 200,000 dwellings included. An unknown proportion of those retirement villages are manufactured housing. 

 

The ABS website explains their methods:

Dwelling location data was recorded by ABS Address Canvassing Officers in the lead up to the 2016 Census as a once-off part of establishing the Address Register as a mail-out frame for designated areas. Dwelling location was also verified or collected by ABS Field Officers during the 2016 and 2021 Census collection periods.

 

In rare cases, an establishment may fall into more than one category of dwelling location, such as a retirement village that contains manufactured homes, or a residential park that is made up of a mixture of caravans and manufactured homes. However, a dwelling can only be allocated to a single category and in these cases a determination was made during Census processing of the most appropriate category for the dwellings in question. 

 

And therein lies the problem, manufactured homes that are not on an estate but within a retirement village. Research on ABS data on retirement villages and manufactured housing estates (MHEs) by Lois Towart found that compared to the 2016 Census “the 2021 Census is significantly more accurate in identifying and recording retirement villages. The issue is the numbers of caravan parks and MHEs that are recorded as retirement villages. This overstates both the size of the sector and the population.” 

 

In her study of 112 retirement villages and 101 caravan parks and MHEs in the Central Coast, Newcastle and Hunter regions in NSW, individual properties were reconciled with small area (SA1) ABS Census data for the 2021 Census. “These are retiree destinations with large numbers of retirement villages and MHEs operated as retirement living’ and “examination of classification by the ABS demonstrates that for the 2021 Census when the dwelling location for caravan parks, MHEs and retirement villages is combined, then the total population is relatively accurate. The inaccuracy is the recording of caravan parks and MHEs as retirement villages.” 

 

How large is this problem? The Census data in Towart’s research has the total number of people living in retirement villages in 2011, 2016 and 2021 as 154,579, then 205,709, and in 2021 249,262. That increase of almost 100,000 people implies at least another 50,000 dwellings, and probably more than 60,000 given the age of this population. An unknown number of those new dwellings were prefabricated. 

 

Some retirement village operators offer sites for relocatable homes, but these will not be classified as MHEs. The construction methods some others use will be based on prefabricated pods and modules, probably sourced locally from a small company. Much of this offsite production might be done by firms not classified as manufacturing buildings. 

 

 

How Many Producers of Prefabricated Buildings Are There?

 

The membership directory of prefabAus lists 9 companies as ‘end-to-end modular’ builders, and there are a dozen other member companies that produce prefabricated buildings or modules. Adding results from other web searches gives a list of 39 companies:

Anchor Homes

Archiblox 

Arkit

Ausco Modular Construction 

Black Diamond Modular Buildings

Carbonlite

Cubehaus

Ecoshelta

Ecoliv buildings

Ehabitat

Fairweather Homes 

Fleetwood Australia

Harwyn

Hickory Group

Hutchinson Builders

Hunter Valley Modular Homes

Habitech Systems

Intermode

K.L. Modular Systems 

Landmark Products 

Maap House

Marathon Modular

Mode Homes

Modscape

Parkwood

Prebuilt Commercial

Pretect

PT Blink

Shawood by Sekisui House

Spanbilt Pty Limited

Strine Environments 

Strongbuild Manufacturing

Sumitomo Forestry Australia

Swan Hill Engineering

Uniplan Group

Valley Workshop 

Volo Modular

XLam

Zen Architects

 

In this (undoubtedly) incomplete list there are substantial companies like Ausco, Modscape and Prebuilt, but many are small firms. Several are engineering and architectural practices that would not be classified as manufacturers. Included are large building contractors like Hickory and Hutchinson, developers like Sekisui and Parkwood, and corporates like Strongbuild and Sumitomo. XLam manufactures cross laminated timber, PT Blink has a design for manufacture and assembly (DfMA) platform. There are also firms that specialise in prefab school buildings, like Marathon, Pretect and Harwyn.

 

Overall it looks like a fragmented market with a few major firms and a large number of small producers, specialised by type of material and type of building produced. Because of transport costs and marketing reach, many producers would be expected to be local and focus on a region. 

 

The diversity of firms in this list highlights the difficulty the ABS would face in measuring the prefabricated building industry. As well as manufacturing, other ANZSIC industries they come from are construction, professional services and business services. With the ABS moving away from surveys to digital data, this sort of detailed data spread across a number of industries is hard to collect. Then there is the question of defining what is prefab and modular construction, which would be needed to organise any data collected and estimate how much is being produced. 

 

 

Deloitte 2023 Industry Survey 

 

Another piece of information comes from the Deloitte Access Economics 2023 State of Digital Adoption in Construction Report, which found 34% of the 229 firms in their survey used prefab and modular construction. The survey sample was from Australia (132), Singapore (38 firms) and Japan (59 firms), and the firms were from Building and construction (144), Architecture (86), Engineering (79) and Other (65). 

 

The survey ranked 16 digital technologies with BIM, construction management cloud software and drones used by around 40% of firm the leaders, followed by prefab and modular where, of the 229 firms, 34% are using it already and 28% are planning to in the future. The remaining 38% are not intending to use it. 

 

That means 76 of the 229 firms use prefab, and its possible that many of them are in the Building and construction category, which would mean up to half of the 144 firms in that category are using prefab and modular. An unknown proportion of those firms are Japanese and Singaporean, where the use of prefab and modular has been supported by government policy and is more extensive than in Australia. 

 

The report included a few more data points:

·      ‘Larger businesses in the survey used significantly more technologies on average, with businesses that have more than 500 employees using an average of seven different technologies, compared with four technologies for those with fewer than 500 employees.’  

·      Businesses used five of the 16 technologies on average, and about 10% were using more than ten different technologies.

·      Newer businesses are investing in new technologies, with ‘businesses less than 10 years old investing 80% more than those in operation for 20 years or more.’

 

The survey does not allow much beyond reinforcing the generalisation that large companies, particularly building contractors, are more likely to be using prefab and modular construction. 

 

 

Conclusion

 

Construction related manufacturing is a significant part of Australian manufacturing, and its share of total manufacturing employment and IVA has increased from 15% in 2014-15 to 16% in 2021-22. There are 17 ANZSIC industry classes included in construction related manufacturing, using data from the Australian Bureau of Statistics annual publication Australian Industry.  

 

The two industry classes for prefabricated buildings are relatively small, but have had rapid recent growth. The share of the combined wood and metal prefabricated building classes in total construction related manufacturing increased by nearly half between 2014-15 and 2021-22. However, wooden building prefab is a very small industry, employing 1,253 people in 2021-21, compared to prefab metal building with 8,144 people employed. The ABS does not have data on what types of buildings are produced (i.e. residential, commercial, mining, institutional etc.). 

 

The ABS 2021 Housing Census found 10,000 manufactured houses in Australia, and another 2,000 townhouses and apartments . This does not include the 200,000 dwellings ABS has in a separate category for Retirement villages. An unknown proportion of those retirement villages are manufactured housing. 

 

The ABS data also does not include offsite work done by firms classified as building or trade contractors, architectural or engineering practices, or work done inhouse in industries like transport, tourism and aged care. The problem the ABS would have measuring the prefabricated building industry is that the ANZSIC industries firms involved come from include manufacturing, construction, professional services and business services, and this sort of detailed data spread across different industries is hard to collect.

 

A list of 39 firms producing prefab and modular buildings was compiled from prefabAUS members and web pages. This appears to be a market with few major firms and a large number of small local producers, specialised by type of material and type of building produced. The industries they come from include engineering and architectural practices, building contractors, and corporates, and any modules or buildings these firms produce will not be found in the ABS manufacturing statistics. This is not a criticism of the ABS, it is an outcome of the classification system used internationally for industry data. 

 

The extent and scale of prefabrication used in Australian construction is unknown, and currently is unknowable. Based on available evidence it is mainly used in specific sectors like the mining industry, education, low cost housing, aged care and retirement villages. There is no evidence that it is cost competitive with conventional construction methods for the great majority of projects. This may be because prefab is a developing industry, or because economies of scale are not as great as expected. 

 

With offsite manufacturing in general, and prefabrication in particular, being seen as important to addressing the industry challenges of sustainability, productivity and skills, this lack of data is regrettable, and the lack of data on how many and what type of prefabricated buildings and components are produced each year in Australia is a problem if an objective of industry policy is to increase the use of prefab and modular construction.