US employers face a decade of 770,000 occupation switches a year, McKinsey projects

Ten years at 770,000 a year is 7.7 million crossings ^[15], about 70% of the 11 million who may have to leave a shrinking occupation ^[16]. The gap fits the research's statement that many of those workers will find new jobs inside the occupational group they already know ^[13]. The 3.6 multiple implies a historical pace of about 214,000 a year ^[17], so the added load is roughly 556,000 workers a year ^[18].

Most of the cost sits in credentials, and the research is looser about them than its own numbers allow. McKinsey describes nearly half the credential problem as a hiring convention and says conventions are easier to change ^[14]. Its figures put 38% of credential requirements down to law ^[7] and another 47% down to employer preferences for a degree, certificate or job title ^[8]. If both are shares of the same base, as their sum to the 85% total suggests, preferences are 55% of the credentialed jobs ^[19]. On that reading the part employers can change is a majority of the problem.

The research does not put a dollar figure on any of this. It does argue that there is a cheaper route than bringing in a credentialed outsider: hire for skills workers can show, remove degree screens the law does not mandate, test candidates with practical assessments and set up apprenticeships for current staff ^[21]. Employers that follow it would spend on internal pathways instead of outside recruiting. Employers that do not risk the case the piece describes: a company that lays off administrative staff in one division while posting project-coordinator openings in another "has created its own talent shortage" ^[20].

The worker's side is harder to budget. For an administrative assistant moving to project manager, the piece says the binding constraint is whether the worker is "willing to invest thousands of dollars and several years to prove it" ^[10]. Forgone wages compound that: someone with limited savings or a child at home may sensibly turn down a move that pays off over 10 years, because it means going months without a paycheck first ^[11]. We think an employer budget that covers tuition but not paid time reaches only the workers who could already afford to stop earning.

The projection could come out smaller. The research frames its main figures as what workers may need to do and what automation could take on, so slower adoption would put crossings nearer 214,000 a year ^[17] than 770,000. Employers that drop preference screens would cut the cost per move, and governments could carry part of the bill through the subsidized education and tax incentives for individuals that the piece lists ^[12]. In our view the line stays in someone's budget under all three outcomes, because its size depends on the credential mix and on wages forgone during training, and both apply at any volume. We would be wrong if measured switching in the first few years runs near the old 214,000 a year.