Better-managed companies are more likely to train their employees, even when compared with firms in the same industries and labor markets, according to October research on how businesses develop skills.
The finding puts work organization alongside training budgets in the debate over how employers prepare people for new technology and responsibilities.
David J. Deming, Raffaella Sadun and Christopher T. Stanton’s research paper combines survey evidence with an economic explanation of why training pays off differently across firms. Circulated by the National Bureau of Economic Research in October, it is also forthcoming in the Journal of Economic Perspectives.
The authors find that training strategies cluster around managerial development or frontline skills. Their explanation is that a company’s structure can make broadly transferable knowledge particularly productive inside that organization.
In their framework, employees who solve more problems themselves free managers to handle harder decisions. That is an explanation of incentives, not a measured return promised to every employer. The survey relationships alone cannot establish that improving management causes more training.
Access to learning remains uneven
The OECD’s Education at a Glance 2026 provides a wider picture. On average, 49% of employed adults participated in non-formal learning, compared with about 34% of unemployed adults and 17% of people outside the labor force.
Non-formal learning means organized education or training that does not lead to a nationally recognized formal qualification. It can include work-related courses, but the participation measure is broader than employer-funded training alone.
The report draws on the Survey of Adult Skills and national sources. Its headline evidence largely describes earlier survey years, including 2023, rather than training delivered in October 2026.
The OECD also reports different priorities by business size. Smaller enterprises were more likely to offer courses on interpersonal skills such as dealing with customers, while medium-sized and large firms placed more emphasis on management skills. That comparison uses European employer survey data from 2020.
These are different measures, populations and reference years. They cannot be combined into a single estimate of what a training program will earn.
A course needs somewhere to be used
For an employer, the research raises a question about what happens after a course finishes. A completion certificate establishes that someone attended or passed an assessment. It does not establish that their daily work now permits them to apply the skill.
Consider a hypothetical customer-service team taught to resolve routine billing problems. If every correction still needs a manager’s approval, the new knowledge may have little effect on response times. Giving staff appropriate authority and a clear escalation process would change the way that training could be used.
That example is an operational interpretation, not an intervention tested in the October paper.
The costs also extend beyond the course fee. Employees need time away from normal duties, colleagues may need to cover their work, and supervisors must be available when unfamiliar situations arise. A cheaper course can become an expensive exercise if none of that time is planned.
We previously examined the mismatch between employees’ abilities and the work they are given. Training adds capabilities; job design determines where those capabilities can be exercised.
For companies assessing a program, measures such as errors, rework, service delays and decisions resolved without escalation can make the intended outcome explicit. Those measures need to suit the job and be compared over a reasonable period, allowing for shifts in demand and staffing.
A business can buy the same course as a competitor. Whether employees get time, authority and suitable work to use it remains a decision inside the company.
