Recall these policy choices:
- Policy 1: one random member of the community is killed, but the rest of us get $101,0112.
- Policy 2: Everyone gets a check for $0.89
To which add:
- Policy 3: status quo
- Policy 4: everyone gets killed
Sartwell values policy 1 less than policy 3 and his value for it is similar to policy 4. In econospeak, he prefers policy 3 to policy 1 and he’s indifferent between policy 1 and 4. In other words, to him just the possibility the government can do harm is as bad as it actually doing harm. So the government of Switzerland, having never caused harm to its citizens but is entity in that territory with the monopoly on violence, is bad because there exists the potential for it to do harm.
Wilkinson is basically pointing out different people may have different policy preferences; they’d tolerate some chance of harm for big enough average benefit.
If arguments about ethics are just about individual preferences, then its not productive to argue over the content of those ethics (at least when talking about optimal policy and in shorter time frames when those preferences are fixed). Instead, the issue for policy is how to aggregate those preferences.