Employer discrimination against smokers, and leveling-down

post by Bill Gardner

Paul restates his arguments against allowing employers to discriminate against smokers (see the comments to this post). His argument, in brief, is that smokers are frequently highly disadvantaged and often have little choice about whether to smoke. Employer discrimination against smokers makes existing class inequities even worse, thereby increasing injustice. As always, Paul writes very well about this. I'm sympathetic but unpersuaded. 

To set up my answer, here are three assumptions:

1. Let's assume that employer preferences for hiring non-smokers are rational, in that they and their firms benefit from them. The benefit occurs because smoking causes workers to be less healthy, and therefore less productive and more costly for employers. 
2. Employer preferences for non-smokers will encourage some but not all current smokers to stop smoking.
3. And let's assume Paul's premise that if this happens, the remaining smokers will be even more disadvantaged. This increases social inequality. Specifically the range and variance of social outcomes increase.

What follows is that employer preferences for non-smokers should not decrease the number of employed. They just mean that non-smokers are more likely to get the jobs. Moreover, these rational preferences make firms more productive, so there is a benefit, averaged across the population. This benefit is increased further by noting that the preferences decrease the rate of smoking, increasing average health. Decreasing the rate of smoking in the current population is likely to further decrease the rate of smoking among the children of current workers. However, assumption 3 concedes that this discrimination increases inequality. 

Conversely, making smoking a protected lifestyle choice would decrease average productivity and health, but it would increase social equality.

If so, preventing employers from discriminating against smokers is a leveling-down policy: something that decreases the variance of outcomes at the cost of lowering the average. There is value in having a more equal society, so if the average is only lowered a little in return for a great increase in equality, I might be in. But I doubt this, particularly when I think about dynamic effects of reduced parental smoking on future generations. We should have strong preferences  for policies that reduce inequality and increase average well-being, like prohibiting racial or gender discrimination.

Smoking is a vice that benefits no one. It's not a candidate for a lifestyle choice that we want to protect through employment laws.

update: Thanks to Paul, who noticed that I wrote "non-smokers" where I meant smokers in assumption 3.