In their simplest and most popular form, consequentialist moral theories identify the morally correct choice as the one that results in the best overall consequences. Despite this criterion's sensible ring, it remains deeply controversial. Many critics object in principle, arguing that a choice may be immoral even though it leads to the best consequences on balance. Here, I explore an alternative possibility, that although consequentialist theories in their simplest form might be attractive in principle, they also might suffer from serious implementation problems.

The difficulty stems from the fact that to engage in consequentialist moral reasoning, one must first construct estimates of the costs and benefits of the relevant alternatives. These estimates almost invariably involve considerable uncertainty, with the result that a broad range of values must be viewed as reasonable. Evidence suggests that even people who are committed to doing the right thing have a natural tendency to exploit moral wriggle room, by employing estimates that favor their own interests. There is a natural tendency, in other words, to estimate the personal benefits of an action at the high end of the reasonable range and to estimate the costs to others of the action at the low end of the reasonable range. Social comparisons reinforce these biases, creating a dynamic that extends the range of estimates that neutral observers can defend as reasonable.