Many of the moral judgments that people term “fairness” seem to implicitly concern distributional effects. Imagine two policies that both affect 10 people. Policy A will cause 9 of the people to gain $10, and 1 to lose $10. Policy A`, an alternative to Policy A, will cause all ten people to gain $5. The average return of Policy A ($8) is greater than that of Policy A` ($5). But many people will say that Policy A is unfair, because, looking at the people who are worse off in each scenario, someone is worst off under Policy A.
This analysis represents a sort of case-by-case, post-hoc Rawlsian difference principle. The analysis is usually applied after the fact of a choice between Policy A and A`. Strangely, perhaps, many people seem not to care who made the choice between the two policies.
For example, imagine ten couples investing their funds for retirement. They choose between one of two investment policies. Under Policy A, they will invest their funds in one of many different sets of stocks. All sets of stocks are expected, with 90% probability, to produce a $10,000 return, and with 10% probability, to produce a $10,000 loss. Alternatively, the couples can choose to invest in Policy A`, which will produce a guaranteed return of $5,000.
Suppose all ten couples freely choose Policy A. We can expect that one of those ten couples will experience a $10,000 loss. People with libertarian sentiments, although they may empathize with that couple, will generally not object, because the couple chose the policy.
However, my impression is that many other people object that this is simply unfair. Even where risks have been chosen voluntarily, they would prefer a more even distribution of losses. Some people will prefer that everyone take Policy A`. They may even support legislation that would force them to do so (such as Social Security).
Whenever you hear an unelaborated claim that some policy or practice is “not fair,” I would suggest that you look for this sort of distributional effect. Does some group of people stand to lose in a big way? Will the unlucky people be worse off under Policy A than under Policy A`?
Some examples of how this “fairness” analysis may color some political debates, high and low:
- Social security may produce a very bad rate of return, but it would be unfair to let retirees lose their money on the stock market.
- Health care insurance may not be worth the cost of the premiums, but it is unfair to force sick people without insurance to pay for the cost of their care.
- Allowing underperforming businesses to fail might aid in the efficient reallocation of labor and capital, but it will be unfair to the unlucky employees of those businesses, who will have to find lower-paying jobs.
- Low skill immigrant workers may increase the output of the economy in general, but they compete with the worst off lowest-skilled Americans and may hurt their wages.
- High credit card fees for missed payments might lower uncertainty for credit card lenders and decrease the total cost of lending, but they will reallocate most of that cost to chronically late payers.
This analysis – unfairness as greater harm for the unlucky or worst off – may strike some people as rather obvious; however it may not be obvious for everyone. And because, I suspect, it is such a central concept, it seems useful to have a go-to post up on my blog for reference.