Some jobs have a peculiar pay structure, in which the guaranteed pay up front may be low, but there can be “exceptional” pay in the future, which may or may not be guaranteed, based on future performance, or on unexpected hardship endured by the employee. For example, consider the typical pension, the ability to get first dibs on overtime by virtue of seniority, or disability benefits in the military.
To act as an incentive for current prospective employees, this payout needs to be out-sized: first, you need to compensate for the time preference of the employee—they would, naturally, rather have the money now rather than in the future—second, you need to compensate for the risk you shift on the employee—it’s rational to prefer a certain future reward to an uncertain one, and so the uncertain reward must be higher in order to influence present behavior.
This creates an inefficiency, in that you have to pay far more than you otherwise would have, in order to incentivize applicants. You also have the second order effects of exceptional circumstances, like that of overtime, becoming routine, further distorting compensation.
After a period of time, you might end up with a pay structure in which a large outlay of your employment expenses goes towards making good on past promises. This creates a massive inefficiency, in that while you push risk onto employees, in the form of uncertain future payment, you yourself have certain current payments to make due on past promises, which are higher than they otherwise would have been had you not moved that risk away. To elaborate further, while the individual may need compensation for holding this risk, because they themselves have a large probability of not benefiting, the firm itself does not need as much compensation, because, averaged over its body of employees, its liabilities are much more certain. Thus, we should want the firm to assume more risk, providing stronger guarantees to its employees, thus reducing unnecessary compensation for risk.
In more concrete terms, it’s much better for firms to set compensation at a level where individuals can decide that the risks of employment—needing to work more during certain periods, and even facing the risk of harm, or death, in the case of the military—are worth the compensation guaranteed to them. This results in both the firm, and the employees being better off.