Does the Risk of Employees Leaving a Firm Affect a Firm's Propensity to Adopt New Technologies?

General-purpose technologies — e.g., electricity, the internet or, more recently, AI and machine learning — spur a process of innovation as they spread through different industries that figure out how to put them to use. “At the core, a lot of that innovation is human capital-related, as it’s people who go through the learning and integration processes and gain experience working with the technology,” says Natarajan Balasubramanian. 

That led the professor of management and two colleagues — Chris Forman, the Peter and Stephanie Nolan Professor of Strategy, Innovation and Technology at Cornell University’s Charles H. Dyson School of Applied Economics and Management, and Ruyu Chen, postdoctoral fellow at the Stanford Digital Economy Lab — to ask in a recent study (under review at Strategic Management Journal) how the availability of human capital affects the adoption of new technologies.  

“Our thesis is that in the early stages of the adoption of a new technology, when a lot of innovation needs to be done and the numbers of workers with the relevant experience are small, these workers are particularly valuable,” Balasubramanian says. “If I think I might lose them to other firms, I might be more hesitant to invest in these projects, and that hinders adoption.” 

The researchers tested their assumptions with a dataset of over 153,000 establishments from 2010 and 2018 and looked at their adoption of enterprise business analytics software. They used changes in state law regarding the enforceability of non-compete agreements — restricting or increasing individuals’ ability to work for a competitor — as a proxy for worker mobility.  

Consistent with their hypothesis, Balasubramanian and his colleagues found that as worker mobility increased, the adoption of the new technology decreased significantly. These negative effects were greater for larger establishments, which — more likely to benefit early on from investments in machine learning technologies — tend to be lead users. The effects were also more pronounced in urban areas, where it is easier for people to find and move to new jobs.  

While previous research focused on the beneficial aspects of worker mobility, this study adds a new dimension by identifying a potential negative effect. “This implies that if you’re a policymaker considering subsidizing adoption of a new technology, you need to be a little more nuanced in your thinking,” Balasubramanian says. “For example, thinking through where will those workers with the relevant experience come from?” This concern, however, is primarily for the short term, he emphasizes. In the long run, worker mobility likely helps spread technological adoption and creative ideas.