Oct. 31, 2024

Can IT substitute for labour at the industry level?

Distinguished Professor Barrie R. Nault, Ph.D., discusses a recent co-authored working paper about the long-time conundrum of investing in IT Capital and impacts on labour costs.
AI and Business
A novel approach looking at industry level data reinforces existing research and discusses emerging concepts

Over decades of emerging information technologies (IT), we have seen that IT and people can work together—sometimes IT can do work for and with people and sometimes IT can also do work instead of people doing it. Accordingly, it is well understood that introducing IT, be it hardware or software, effects costs of labour. Emerging research [1] by BTMA Distinguished Professor, Dr. Barrie R. Nault and co-authoring colleagues [2] draws on essential IT theory and is groundbreaking: the first evidence showing the impact of IT Capital on transaction labour costs at the industry level.

"To some of us, Transaction Cost Theory for which Oliver Williamson was awarded a Nobel Prize, is a theory that can explain many things, but transaction costs are notoriously hard to measure. We follow an approach by Douglas C. North, another Nobel Laureate who is also considered an intellectual father of this years’ Nobel Awardees, along with his co-author John J. Wallis that measures transaction costs through labour costs associated with certain industries and occupations within industries."

Moreover, the research team examines how IT can substitute for or complement labour in the industry setting by looking at how IT impacts two types of labour costs: transaction labour costs such as those related to supervision of production, human resources, accounting and other management costs (internal costs) or coordinating outside purchases, contracts (external costs), for example; or transformation labour costs—all costs associated with production excluding materials and other intermediate input costs.

"The focus of this research is the effects of IT on costs related to transaction labor relative to transformation labour. That is, our approach examines the relative changes in transaction versus transformation labour costs due to changes in IT and non-IT capital at the industry level—this is what is novel about our approach."

Dr. Barrie R. Nault, Distinguished Professor

Dr. Barrie R. Nault, Distinguished Professor

These affects are measured across all industries in the US economy over a 13-year period (2009-2021). Using econometric methods, a cost-function based analysis and a production-based analysis, the research team interprets data including COVD-19 years, which had similar results to non-COVID-19 years. 

"It is surprising that the COVID-19 years did not yield anything different. This probably speaks to the robustness of the economy and that the hit in the employment part of the economy was pretty evenly spread across occupations. Our analysis is economy-wide, we include all industries – so there is no different set of industries. This makes our results all the more powerful."

Findings also show that IT has substituted for transaction labour to a greater extent than transformation labour based on task type: routine and non-routine cognitive tasks.

"Indeed – labour roles are effectively determined by occupation—IT is more likely to substitute for transaction occupations and complement transformation occupations. Furthermore, IT has had a greater impact on routine tasks rather than non-routine tasks."

Although inherently speculative in theory, the research findings help paint a picture of how IT with emerging AI capabilities might impact labour costs in industries, 

"The theory, routine-biased technological change or RBTC, suggests that IT increasingly substitutes for labour that performs routine tasks. We find that the transaction labor cost share allocated to routine tasks declines with increased IT, and I would expect that AI will accentuate the impact of IT on labour costs for routine tasks. Whereas non-routine tasks are more akin to “exceptions”, and I suspect AI will have a more difficult time substituting for humans there."

Accordingly, findings suggest this deepened relationship between IT and labour, whereby “innovations in IT together with declining prices of IT, have replaced workers performing routine tasks with machines.” [3]

Drawing on relations to previous research, Nault further explains how the type of IT being used matters when measuring the extent that labour costs impact both markets through firm structures, 

"It really depends on the IT we are talking about – we have other research suggesting that IT hardware, which would include production machinery such as CAD/CAM, CNC machines, etc., reduces internal transaction costs favoring “make” versus “buy”. On the other hand, IT software such as enterprise systems and supply chain management systems reduce the cost of external procurement, favoring “buy” versus “make”. "

This nuance for firm structures could be a future development of the research. 

_________________________________

[1] Han, K., Koh, B., Nault, B.R., & Park, J. (December 2023). The differential impact of IT on transaction and transformation labor. Working Paper

[2] Associate Professor Dr. Kunsoo Han, McGill University; Professor Dr. Byungwan Koh, Korea University; and Assistant Professor Dr. Jiyong Park, University of Georgia

[3]Han et al. (December 2023), page 34

Find more information about iRC research and activities here.