Construction Productivity Trends for Building, Engineering and Construction Services

 Australian Construction Productivity at the Industry Level

 

 

The rate of growth of productivity in the construction industry in a number of countries has lagged that of other industries for at least five decades, and the earliest studies that identified this problem date from the late 1960s. Two explanations for the lack of demonstrable improvement in construction productivity are possible. The first is the importance of measurement, data and issues about the structure and use of price indices for estimating real output (i.e. adjusted for inflation). 

 

The second is the nature of the product and the methods used in delivering and managing the processes involved.  Construction is a labour intensive industry in comparison with manufacturing, but there has been a significant increase in the prefabricated component of construction, which could have been expected to lead to productivity growth. Also, construction methods have become more capital intensive as machinery has got heavier, and the number of cranes, powered hand tools and other equipment used has increased.  However the productivity growth that one would expect to observe as a result of these trends has not occurred, according to measurements by national statistical agencies.

 

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 that assume there has been no change in the quality of buildings. Another problem is the application of a single deflator to the diverse range of buildings and structures. This inability to capture functional differences and quality changes in buildings and structures has adversely affected the measurement of productivity, if construction value added is underestimated due to the deflators used, construction productivity has also been understated.

 

This post compares the deflated GVA per person employed to the IVA per person employed for Building, Engineering and Construction services (the trades), and Total construction. The GVA data comes from the ABS National Accounts (chain volume measures of economic activity). The IVA data and number of people employed in June each year comes from ABS Australian Industry. 

 

 

A Proxy for Construction Productivity

 

In Figure 1 industry output is in constant dollars (the deflated value adjusted for price changes). GVA is the quantity of output produced in a year. The employment data includes all workers but not whether they are full or part-time, or hours worked. 

 

Figure 1. Construction Productivity by Industry

Source: ABS, CER

 

As a measure of productivity GVA per person employed is very approximate, typically the number of hours worked would be used for employment and June may not be a representative month for employment in many industries. Nevertheless, this graph looks familiar, with flatlining growth in Total construction productivity over the period, despite a few bumps along the way. It appears to be a useful productivity proxy. 

 

Using the same data, GVA per person employed can be found for Building, Engineering and Construction services. Here a slight decline in Building has been offset by a small rise in Construction services output per person, with the effects of the pandemic on both apparent in the decline over 2020-21. Building construction may have been affected by a shift from commercial to an increased share of residential in the output mix and more high rise work. Because Construction services are generally labour intensive they will have a lower value of output per person, but this data shows there was increase in this measure of productivity between 2007 and 2021 and Construction services was the only one of the three industries to register a gain on this measure. 

 

Engineering construction activity took off in the mining boom from 2010, and output per person has followed the rise and fall in work done since and, although below the peak years of 2012-14, it now reflects the large volume of infrastructure work in transport and energy. Since 2011 GVA per person in Engineering has been much higher than Building construction, nearly twice as much in some years, and Construction services, nearly three times as much in some years. 

 

These differences in output per person employed reflect differences in capital requirements and expenditure on purchases of buildings, structures, software, equipment and machinery (known as gross fixed capital formation or GFCF). 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.

 

 

Current Dollar Industry Comparison 

 

The chain volume measure of GVA per person employed can be compared to the original, unadjusted current dollar Industry value added (IVA) per person employed. Again, this is an indicative but imprecise proxy for construction productivity. In Figure 2 there is a clear upward trend in all three industries, with increasing nominal value of output as prices rise faster than the number of people employed. 

 

The growth in IVA per employee for Building is the greatest contrast to the GVA data. Here, Building has had a sustained increase since 2012 compared to the flat, no growth trend in GVA per employee. This suggests there has been a better productivity performance by building contractors than the one recorded in official statistics. 

 

Engineering has a similar pattern in both GVA and IVA graphs, with a sharp rise in output per employee after 2010 that flattened out after 2016 at around 50 per cent higher than the pre-mining boom level. This has been a significant increase in productivity. Both Building and Engineering typically have larger firms than found in Construction services, which has lagged the other two industries in growth in IVA per employee. 

 

Without deflation the value of output could be expected to rise somewhere around the rate of CPI inflation, which totalled 35.8 per cent and averaged 2.2 per cent a year between 2007 and 2021. Over that period Building IVA increased by 120 per cent, Engineering IVA by 117 per cent, and Construction services by 50 per cent. More significantly, IVA per person employed for Building increased by 61.6 per cent, for Engineering by 57.3 per cent, but for Construction services only 27.7 percent, suggesting that is where the productivity ‘problem’ lies. However, the IVA and GVA figures are contradictory, with the latter showing better performance. 

 

 

Figure 2. Nominal output per employee

Source: ABS, CER

 

IVA per employee again highlights differences in the capital requirements of industries. In the long run, investment in GFCF determines industry growth rates and their level of labour productivity. Labour intensive industries like Construction services have a low level of IVA per person employed, but also have lower capital requirements. Engineering has always been more capital intensive than Building, but the gap seems to have closed with the increase in residential high-rise activity after 2016. 

 

Conclusion

 

Construction productivity estimates are usually given for Total construction, and typically show little or no growth over many decades. However, Total construction is measure of the combined performance of three different industries: Building, Engineering and Construction services. This post compared the deflated GVA per person employed to the nominal IVA per person employed for Building, Engineering and Construction services (the trades), and Total construction.

 

The deflated GVA per person employed data is a proxy for productivity because the value of output is adjusted for price changes, As a combination of deflated output and employment GVA per person employed looks like a measure of productivity, but while it is indicative that is not really the case. Although similar to the output and input data needed to calculate productivity, indexes of output and input are used for productivity analysis, not the original data, and hours worked not numbers employed used. 

 

When the mostly flat chain volume measures of GVA per person employed are compared to the current dollar IVA per person employed there is a clear upward trend in IVA all three industries, with increasing nominal value of output as prices rise faster than the number of people employed. IVA per person in Building and Engineering has increased at nearly twice the rate of CPI inflation, but Construction services by less since 2007. 

 

Construction services IVA per person employed grew significantly less than Building and Engineering. However, the GVA per person employed performance was much better, the only one of the three industries to register a gain on this measure. Construction services have a large impact on productivity because they account for 60 per cent or more of Construction output. 

 

The usefulness of both GVA and IVA per person employed as a proxy for productivity per person is limited, but indicative. In both cases the difference in capital intensity appears to be the determining factor in the level of productivity (measured as dollars per person employed), and the increase in apartment building would explain the rapid rise in Building IVA per person employed. The effect of changes in output (the mix of buildings and structures delivered) will be explored in another post. Why that increase in Building IVA per person employed was not picked up in the GVA per person employed estimates is also an interesting question. 

New US Data Shows We Have a False Picture of Construction Productivity

 Rises in Non-residential and Declines in Residential Construction Labour Productivity

 

 

The tools, techniques, components and materials used in modern construction can be seen on every building site. As anyone who works in construction knows, they have greatly increased the productivity of workers, but that increase in productivity cannot be seen in construction statistics. For decades there has been little or no growth in construction productivity as measured by national statistical agencies. 

A major problem is the inclusion in Construction sector statistics of residential building, non-residential building and engineering construction. These are three separate industries with significant differences in their characteristics, but statistical classifications group them together despite these differences.  For example, non-residential construction uses heavy machinery and equipment and is much more capital intensive than residential building. Most construction productivity research uses this aggregate data for construction because in the statistics published by national statistical agencies data on employment and hours worked at the level of the three industries is missing. 

Productivity estimates require both a measure of labour inputs, such as hours worked or people employed, and a measure of output, usually industry value added (the difference between total revenue and total costs) adjusted for changes in prices of materials and labour. That deflated measure of output is known as real construction value added.

 

The main reason for the low rate of measured productivity growth in construction are the deficiencies found in construction deflators. If real construction value added is underestimated due to the deflators used, construction productivity has also been understated. Thus the graphs of flatlining construction productivity, despite the obvious improvements in materials, tools and techniques over the last few decades. 

 

The major problem is a downward bias to output estimates because there is no adjustment for quality changes in buildings and structures.  Also, the application of a single deflator to heterogeneous goods, especially durable goods, overlooks the differences in age and function between different buildings and structures. This problem becomes more severe with long-life assets like buildings and structures. 

 

 

New Data on US Construction Productivity

 

Three economists at the US Bureau of Labour Statistics recently published productivity estimates for fourconstruction sub-industries using four different deflators, providing new, high quality estimates of real construction value added per hour worked in these industries, including subcontractor hours. The BLS research improves on previous research by using appropriate output deflators to develop measures of productivity growth,  and their measures are more reliable because the deflators are specifically designed for each industry. Their data and analysis is a significant advance on the aggregate construction productivity estimates that people are familiar with. The four industries are:

• industrial building construction
• single-family residential construction
• multiple-family housing construction
• highway, street, and bridge construction 

As the figure below shows productivity fell in single-family residential and multiple-family housing construction, but rose in industrial and highway, street, and bridge construction. Between 2007 and 2020 overall productivity was flat because these rises and falls balanced out. Also, 2007 was the peak of a business cycle, followed by a recession from December 2007 to June 2009 that ‘had both immediate and lasting impacts on the construction industries’ as the following figures show.

Two of the four industries show clear and strong productivity growth. Productivity growth in these industries remains positive with subcontractor labour included. Productivity grew fastest in industrial building construction

 

Labour Productivity in US Construction

The BLS figures below show trends in output, hours worked, and labour productivity for each of the four industries. Importantly, ‘labour hours always include partners and proprietors (P&Ps), who account for almost 20 percent of labour input in construction.’ Because of data limitations, the reference periods of the industries begin at different points. The BLS comments are under each figure. A short discussion on construction deflators follows. 

 

Single-family residential construction

For single-family residential construction labour productivity rose during the 2000-2005 period, primarily driven by a large increase in output. Starting in 2005, output fell through 2009 at a considerably faster rate than hours worked, leading to a sharp decrease in productivity in the period. These years correspond with the collapse of the housing market. Labour productivity grew from 2009 to 2013 but steadily weakened through 2019.

 

Multiple-family housing construction

Multiple-family housing had large gains in productivity from 1993 to 2007, as output increased faster than hours worked, followed by a sharp decrease in output and a moderate drop in hours worked, leading to decreasing productivity until 2010. From 2010 to 2016, output rebounded substantially, leading to significant productivity growth. However, productivity dipped from 2017 to 2020 due to growth in hours worked outpacing growth in output

 

Industrial building construction

For the industrial building construction industry, from 2006 to 2018 the productivity and output indexes rose until 2009, then fell sharply in the 2007-2009 recession, recovered from 2011 to 2015, and then declined again through 2018. The change in hours worked was slow and uneven, but positive over the 2006-2018 period. In 2019, the rising output and falling hours worked series moved in opposite directions, which led to the first gain in productivity since 2015.

 

Highway, street, and bridge construction

Productivity in the highway, street, and bridge construction industry increased during the 2007-2009 recession. Output rose as hours worked declined during this period. Then, until 2018, productivity fell most years as output saw no net growth while hours worked did. Since 2018, there has been little change to output, hours worked, or productivity.

 

Construction Productivity

Construction productivity has been notable for its absence for decades. The low rate of growth in Australia, the United Kingdom, the European Union and the United States and elsewhere became an issue in the late 1960s, when declining output per hour worked and output per person employed in the construction industry first attracted attention. The measured rate of growth of productivity of the construction industry since then has been poor even by comparison with a long-run overall industry average in the order of two to three per cent a year. Construction productivity in the US has been falling since the late 1970s, as in the figure below.

Possible reasons for the low growth of construction productivity are many and various. They include the high labour intensity of the residential industry, the number of small firms, few economies of scale in the industry, a lack of competition, regulatory impediments, low R&D, poor innovation and management practices, and a low level of training and skill development.  Alternatively, it is possible that the data and methods used to estimate the level of productivity in the industry might be faulty, and we have a false picture of construction productivity.

A second and more technical problem is the method used to adjust industry output for changes in prices of materials and labour to find changes in the quantity of output of completed buildings and structures. This deflation of output is typically done with input price indexes or producer price indexes. The problem with input price indexes is they assume a constant relationship between input and output over time, so there is an assumption of no change in productivity which means that, if productivity is increasing, input price indexes will be upwardly biased.

There is an extensive literature on deflators, the problems of deflation, and the effects on estimates of construction output of commonly used deflators. The issues raised by the use of price indexes for deflation have not been solved to date, and appear to have no simple, or readily available solutions.  These include the fact that the deflator used to adjust for price changes will systematically overstate the rate at which prices increase and underestimate growth in output if indices for labour and material costs are used instead of output price indices (which are generally not available). I have a 2001 paper on Construction Deflators and Measurement of Output. 

 

Another problem is the application of a deflators to the diverse range of buildings and structures, and differences in quality and function between them. As the energy efficiency and quality of finishes has improved, and as the share of building costs due to mechanical and electrical services has increased over time (providing greater amenity), the deflators used have not been adjusted to take these trends into account.  In effect, the deflators assume there has been no change in the quality of buildings, and their inability to capture quality changes in the buildings and structures delivered by the construction industry has adversely affected the measurement of productivity.  

 

The US Bureau of Labour Statistics recently published productivity estimates for four construction sub-industries used four deflators from different government databases. Their research addresses the problem with new data: ‘The main difficulty is that buildings differ widely in their characteristics and features. Similarly, the nature of the underlying terrain varies widely among construction projects. Consequently, economists, both in general and within the BLS productivity program, have found it exceptionally difficult to develop reliable output price deflators to convert observed revenues into meaningful measures of output growth over time. Good output price deflators are therefore the key to more accurate measures of productivity growth in construction.’

 

The researchers say: ‘we examine only those industries in which the deflators exactly match the industry boundaries. Previous work generally looked at the total construction sector. Since the many new deflators now available did not exist then, these prior studies had to use the single-family housing deflator and an associated cost index to deflate production in most or all of construction.’

 

 

 

Leo Sveikauskas, Samuel Rowe, James D. Mildenberger, Jennifer Price, and Arthur Young, "Measuring productivity growth in construction," Monthly Labor Review, U.S. Bureau of Labor Statistics, January 2018, https://doi.org/10.21916/mlr.2018.1.