Never Having to Risk It
The Office: S3, E20 “Safety Training”
Labor Markets: Compensating Differentials
Darryl gives an eye-opening safety demonstration in the warehouse. But when Toby leads the safety meeting in the office upstairs, the dangers don’t compare. He mentions carpal tunnel syndrome, the strain of a computer screen on the eyes, and the possibility of feeling a bit cold. So, when Michael tries to compare what happens in the office to what happens in the warehouse, Darryl won’t hear it.
Darryl: “That's what I've been trying to tell you, Mike. It's serious down there. We do dangerous stuff, man. This is shenanigans, foolishness, Nerf-ball. You live a sweet, little, nerfy life. Sittin' on your biscuit. Never havin' to risk it.”
Michael takes this as a challenge: prove to Darryl that office work is just as dangerous as warehouse work. He plans to emphasize the risk of depression and puts together a demonstration for the warehouse workers. But oddly, he also invites his office workers.
The presence of the office workers matters, because all else equal, workers get paid more as their jobs become more risky, dangerous, or otherwise less appealing. But this is only true to the extent they are aware and respond to the dangers.
To see why, consider the supply and demand for office work specifically at Dunder Mifflin.1 As the salary increases, more people want to work at Dunder Mifflin. But because of diminishing marginal product of labor, Dunder Mifflin wants to hire fewer workers as the salary increases. How are the salary and the size of the staff determined? There is an equilibrium salary that makes the number of applicants equal to the number of workers Dunder Mifflin would want.2
Suppose Michael’s demonstration works. His workers now believe that their nerfy jobs will induce depression in no time. Their changed outlook would reduce supply, since the supply curve quantifies how many people would want to work at Dunder Mifflin for any given salary.3 If every hire were given a free car, supply would increase. And if every hire assumed they would become seasonally depressed, supply would decrease.4
The consequences of Michael’s mismanagement are best seen graphically.
Dunder Mifflin corporate will not be happy with this blunder. The decreased supply leads to higher equilibrium pay: a $4,000 raise for every worker.5 This difference in pay is the risk premium associated with the exaggerated belief that this line of work is depressing and dangerous. Workers who take on jobs that (they believe) are worse for non-salary reasons are compensated with higher salaries. These are compensating differentials.
Isn’t this good news for the office workers? Maybe they should hold their own safety demonstrations and convince themselves that the dangers are even greater than Michael claims. But the law of demand pinches that logic. Once Dunder Mifflin must pay higher salaries to induce enough workers to show up, it wants to hire fewer of them. But why does the financially constrained paper company raise pay at all?
Holding pay fixed at its previous level would leave Dunder Mifflin wanting a 15-worker staff with only 10 workers willing to offer their time and energy. It’s forced to raise the salary to induce more workers to show up.6 But crucially, this change won’t lead to a salary big enough to convince every worker to stay. At $54,000, only 14 workers are willing to stick around. The 15th is not compensated enough for their perceived depression risk. So, this turns out well for 14 office workers, but goes poorly for one worker and the employer. It’s clear Michael has created another mess for Dunder Mifflin: a smaller, higher-paid staff.7
This story has implications for the real world as well. From a policy perspective, how should dangerous jobs be addressed? Let’s focus on cases where the danger is quite real. Underground miners suffer from poor air quality. Machinists lose their hearing over time. Football players suffer from head injuries. Rodeo clowns can be gored. To some extent, these workers are compensated for the risks they take on. But that only happens if they are aware of the risks! And employers have little incentive to uncover risks and communicate them to their employees. They may even hide risks they become aware of. Most bosses are not like Michael Scott.
So, governments can solve real problems by researching and monitoring job safety to provide useful information to workers.8 Information provision allows workers to be adequately compensated for the real risks they take. And once risk premiums are factored in, employers have incentives to create safer workplaces.9
The government can probably save its resources inspecting Dunder Mifflin Scranton. Its boss has a habit of revealing the dangers to his workers, even dangers that don’t exist or have already been cleared. He does so again not long after his depression demonstration.
Michael: “I wanna get out of here. All the cliques and the office politics. Fluorescent lights. Asbestos.”
Jim: “I thought we had that looked at.”
1 This is a generalization across many finer types of work: accounting, sales, etc. The same steps used here could be used within each of these specific types of work instead and the same conclusion would be reached.
2 This is a simplification. If the supply of workers to Dunder Mifflin specifically is upward sloping, then Dunder Mifflin has some monopsony power, and the salary that results is not one that equates quantity supplied and quantity demanded. A richer model is needed. But the richer model changes the level of salaries, not the logic that follows about workers’ knowledge of dangers.
3 The supply curve also reflects potential office workers. We may assume that they are similarly affected by the demonstration, despite not seeing it. Perhaps word gets around town.
4 The height of the supply curve also reflects reservation salaries of marginal workers. For workers who are bound to office work regardless of employer, this news may have no effect on their reservation salary to work at Dunder Mifflin. The other places they could work are similarly depressing, surely. But some will increase their reservation salaries, because they could realistically get a job without the danger. This logic leads to a supply shift that is not parallel but rotates the supply curve up instead.
5 What determines the magnitude of this change in pay? Within this model, elasticities of supply and demand are primary determinants as well as how much supply decreases. The type of model matters too. A more realistic monopsony model would have a different prediction of the magnitude of the risk premium.
6 Suppose you’re thinking “There’s no way a third of the staff will credibly threaten to quit unless their pay goes up.” That’s possible and would mean that the supply shift post-demonstration is smaller than I have modeled it. Therefore, the risk premium would be smaller. The risk premium scales with the magnitude of the supply response.
7 From a welfare perspective, there is a deadweight loss as well. The 15th employment contract was mutually beneficial before the exaggerated risks were conveyed. It still would be. The workers are reacting to a false belief. Here, the drop in employment creates the deadweight loss. But note that in a scenario where workers are unaware of a real danger (like asbestos in the workplace), an employment drop due to workers responding to accurate information about the danger removes a deadweight loss.
8 Don’t workers already have incentives to do this? Not necessarily. Each football player hopes the others are looking into the long-term health risks. Here, information is a public good. And public goods are often underproduced due to the free rider problem.
9 Why not just mandate safety? Without mandates, firms will invest in safety whenever the wage-bill savings are greater than the investment costs. With mandates, firms will at times be forced to buy safety out of workers’ wages when workers would prefer to keep the higher salary and assume the risk.
More from The Office:
1 This is a generalization across many finer types of work: accounting, sales, etc. The same steps used here could be used within each of these specific types of work instead and the same conclusion would be reached.
2 This is a simplification. If the supply of workers to Dunder Mifflin specifically is upward sloping, then Dunder Mifflin has some monopsony power, and the salary that results is not one that equates quantity supplied and quantity demanded. A richer model is needed. But the richer model changes the level of salaries, not the logic that follows about workers’ knowledge of dangers.
3 The supply curve also reflects potential office workers. We may assume that they are similarly affected by the demonstration, despite not seeing it. Perhaps word gets around town.
4 The height of the supply curve also reflects reservation salaries of marginal workers. For workers who are bound to office work regardless of employer, this news may have no effect on their reservation salary to work at Dunder Mifflin. The other places they could work are similarly depressing, surely. But some will increase their reservation salaries, because they could realistically get a job without the danger. This logic leads to a supply shift that is not parallel but rotates the supply curve up instead.
5 What determines the magnitude of this change in pay? Within this model, elasticities of supply and demand are primary determinants as well as how much supply decreases. The type of model matters too. A more realistic monopsony model would have a different prediction of the magnitude of the risk premium.
6 Suppose you’re thinking “There’s no way a third of the staff will credibly threaten to quit unless their pay goes up.” That’s possible and would mean that the supply shift post-demonstration is smaller than I have modeled it. Therefore, the risk premium would be smaller. The risk premium scales with the magnitude of the supply response.
7 From a welfare perspective, there is a deadweight loss as well. The 15th employment contract was mutually beneficial before the exaggerated risks were conveyed. It still would be. The workers are reacting to a false belief. Here, the drop in employment creates the deadweight loss. But note that in a scenario where workers are unaware of a real danger (like asbestos in the workplace), an employment drop due to workers responding to accurate information about the danger removes a deadweight loss.
8 Don’t workers already have incentives to do this? Not necessarily. Each football player hopes the others are looking into the long-term health risks. Here, information is a public good. And public goods are often underproduced due to the free rider problem.
9 Why not just mandate safety? Without mandates, firms will invest in safety whenever the wage-bill savings are greater than the investment costs. With mandates, firms will at times be forced to buy safety out of workers’ wages when workers would prefer to keep the higher salary and assume the risk.