The problem with the theory around mid-range-λ-jobs is that people with a non-zero t can compete for jobs as well - for example, they may have increased their s

The problem with the theory around mid-range-λ-jobs is that people with a non-zero t can compete for jobs as well - for example, they may have increased their skill level through training, volunteer work, formal education, or there could be a flood of skilled workers following layoffs at a competitor. A skilled employee is therefore potentially competing both against 'green' replacements and skilled replacements. At times when there is a surplus of skilled employees, the market price for skilled employees will tend towards the marginal cost of living for skilled employees, which is roughly the same as the costs of new employees, and so new employees would be prevented from gaining experience on the job (they could offer to work for less than their cost of living in the hope of a future payoff from their skills, but even that won't necessarily make sense to the employer, because the employer has costs per employee beyond just wages and employee benefits, and it might still be cheaper to hire fewer skilled employees at a higher price than to incur more per-employee costs hiring more unskilled employees at a lower rate). On the job learning therefore becomes less relevant, and new employees are forced to overcome the activation costs through other means, or try a different industry. If there is no credible alternative to on the job learning, this might cause a dynamic effect where the attrition of skilled employees combined with an absence of trained new employees raises the skilled employee price towards the utility to the employer, and makes it cost-effective to take on new employees again.