By Diane Coyle and Rehema Msulwa
Construction is a large sector of the UK economy, while investment in infrastructure is often seen as an important contributor to productivity. The National Infrastructure Commission in its baseline report for the next national Assessment has highlighted the need for substantial investment in several areas of infrastructure: transport connectivity, water and sewage systems, flood management and decarbonization.
Yet official statistics show that productivity in this area of construction has flatlined despite the increasing adoption by the sector of advances in digital technology and increased outsourcing. This is consistent with the popular image of construction sites as low-tech environments. Labour productivity in infrastructure in the UK is about the same now as in 1997, even though some parts of the supply chain are now extensively using digital tools.
In a new ESCoE Discussion Paper, we examine the evolution of the infrastructure construction industry and the scope for its digitization, describing the changing industry structure and the range of digital technologies being adopted. We consider the implications of technological innovation for productivity, both actual and measured. We focus on two possible explanations for this particular piece of the productivity puzzle: time lags and different potential sources of mismeasurement.
In the paper, we explore the impact of digitization on the structure and productivity of the UK’s civil construction sector with a focus on major projects. In the infrastructure sub-sector there has been substantial change in market structure since 1980, along with accelerating technical change. In addition, designs and materials used have seen innovation, particularly to bring about environmental improvements in construction and use. The use of building information modelling (BIM) tools has steadily commoditized the design stage, leading to considerable outsourcing to countries such as Brazil, India and Poland. On sites, much construction activity is contracted out to specialist sub-contractors. In between, both consultants and major contracting firms provide digitally-enabled project and programme management using integrated 3D software models that capture data through time from design through construction to operation (sometimes known as ‘digital twins’). For example, sensors embedded in concrete or steel can be monitored in real time allowing reduced maintenance costs or averting structural failures. In addition, parts of large structures are increasingly built offsite for onsite assembly, a process enabled by BIM, reducing construction time and improving component standardization and quality. Some major firms are building vertically-integrated offsite factories to produce these parts for their construction projects. Finally, the incorporation of digital sensors is enabling provision of related maintenance services following completion of a project.
While all these developments are widely recognized in the industry, these changes are seemingly not yet well reflected in output, price and productivity statistics. This could be due either to only gradual implementation of the technologies across the sector, or to measurement issues such as failure to quality-adjust deflators (although there may well be other reasons for productivity headwinds, such as planning restrictions, that we don’t discuss).The former, often described as a productivity J-curve effect, would imply simply delay in seeing the productivity benefits of digitzation. Issues with output or input price deflators, which would imply some mismeasurement of productivity, could arise from the absence of relevant quality adjustments (eg enhanced asset lives or performance), or from the failure to capture some prices (eg outsourced services). There have also been activity shifts across SIC 2007 classifications, which could have led to mismeasurement of relevant input volumes and deflators.
Part of the explanation for the failure of the sector to realize productivity benefits from digital tools may, however, lie in the lack of coordination along the value chain. An important defining characteristic of the industry today is its organizational separation from clients. This contrasts with some other industries where continuing services are becoming a more significant part of value creation in the ‘servitization’ process, although this is also starting to change with infrastructure. On the whole, though, contractors and their supply chain tend to have limited involvement with clients upfront in the feasibility stage of a project. The lack of integration along (sometimes global) supply chains is manifested in a wide-scale use of on-site sub-contracting and multiple tiered interfaces and significant transaction costs. These include multiple overhead costs in contracting and co-ordination, and (often inappropriate) risk transfer to sub-contractors. This structure in turn has led to an industry that tends to be focused solely on cutting costs rather than adding value.
As a result, the industry delivers bespoke products and is unable to deliver process efficiencies. Vertical fragmentation occurs because each project stage requires a different set of stakeholders, decision-makers and values. The decoupling of different stages can give rise to self-interested behaviour and project participants passing costs on to participants in subsequent stages. Longitudinal fragmentation occurs because project teams disband at the end of individual projects and are selected on future projects by fresh competitive bidding. Taken together, this means that there are significant transaction costs in all infrastructure projects, with substantial fragmentation of work on-site in particular among the many subcontractors. This makes it clear that – as in many other sectors – there is likely to be a “productivity J curve” (Brynjolfsson, Rock & Syverson 2021) or delay in adoption because of the need for organizational change – including data standards and data sharing.
The paper also considers whether there are measurement challenges related to construction deflators – either the prices paid for inputs such as (now often outsourced) design or the output prices for what are often distinctive projects. Industry statistics suggest that revenues per hour vary enormously even across projects in the same category such as rail or road schemes. Both the inherent measurement challenges and the J-curve seem likely to contribute to explanations for slow productivity gains (along with factors we did not consider, such as planning), despite the increasing digitization of infrastructure construction.
Brynjolfsson, E., Rock, D., & Syverson, C. (2021), ‘The Productivity J-Curve: How Intangibles Complement General Purpose Technologies’, American Economic Journal: Macroeconomics, 13(1), 333-72.
Read the full ESCoE Discussion Paper here.
ESCoE blogs are published to further debate. Any views expressed are solely those of the author(s) and so cannot be taken to represent those of the ESCoE, its partner institutions or the Office for National Statistics.