WPC Research
“Our framework could help assess how this social skill premium makes people climb the ladder faster.”
Research by Andreas Kostøl, Assistant Professor of Economics
“Our framework could help assess how this social skill premium makes people climb the ladder faster.”

Inside job: Understanding internal labor markets


e’re going through a period of unprecedented changes in the employment market, which makes it an excellent time to take a closer look. We usually think of job markets as being a space between companies, where potential employees and employers go to connect. But there’s another kind of labor market — the internal labor market (ILM), which exists within a given company.

“Economists don’t have many models for ILMs,” says Andreas Kostøl, assistant professor of economics. “There is no clear sense of what they look like and how they work.”

That’s why Kostøl and an international team of economists developed a model that sheds new light on ILMs.

Data in demand

The first thing the researchers needed was data, which they found in Norway, where the law requires companies to report all employee status changes. Kostøl is Norwegian, and his team represents scholars from Vienna, Oslo, Berlin, and Tempe, Arizona.

With access to comprehensive data, they needed an algorithm to analyze it — specifically, they needed to infer the ranking of job positions within a firm. Often, job titles are too vague to determine the relative value of one position compared with another. So the team developed an algorithm that determined this based on the amount of employee traffic a particular position was attracting in relation to another: The more traffic to the job, the higher the value.

This is the first time economists have had an automated methodology for analyzing big data sets on ILMs and their hierarchies. Using this methodology, the team could quantify the existence of ILMs and identify their characteristics.

Variety is the constant

They found ILMs vary dramatically in their structure and hierarchy. Even companies of similar sizes and scope of operations can have very different ILMS. The researchers were also surprised to learn that multiple ILMs can exist within a single company.

“We saw that the core business operations can form one ILM,“ says Kostøl, “and people who work in support services, like janitorial crews and receptionists, might constitute a separate ILM within the same company.”

This finding could significantly affect how we understand and measure wage and promotion opportunities within companies.

Examining gender inequality and premiums on social skills

“The main innovation is the algorithm that lets us see these hierarchies within ILMs,” says Kostøl. “In the future, we can apply this to clarify other questions.”

One area he wants to explore is the inequalities women might face in climbing the same job ladder as men. “Are women facing different challenges than men?” he asks. “This methodology can help quantify that.”

Another question that interests him is the role of noncognitive skills in determining pay. Social skills are becoming more important and more highly prized in the workplace, but what is the reason for this? Is it because companies are working more in teams, giving more value to people who can cooperate and collaborate? Or is this trend driven by the need for more salespeople or more roles to take on a sales function?

“Our framework could help assess how this social skill premium makes people climb the ladder faster,” says Kostøl.

Big data could bring big clarity to HR

This work is part of a more significant trend to apply big data to human resources. Big data is already transforming many business areas, including marketing, manufacturing, and transportation, “but HR has not yet been affected — until now,” Kostøl says, adding that Amazon recently hired a data scientist to reorganize its HR strategies around big data.

In addition, business schools are starting to add programs in related subjects such as HR analytics, which is good timing for Kostøl and his colleagues’ work.

— Joe Bardin