A person gets cancer. There is a therapy that has a good shot at keeping her alive, but it’s expensive. After fighting with insurance, she does get the therapy, but months have passed. The cancer has advanced and it’s too late. She dies. 1/

She might still have died had she received the therapy promptly. Or not. No one knows, but her odds would objectively have been much better. 2/

In medical consumption terms, she’s profoundly richer than a doppelgänger in her situation 20 years ago. The therapy didn’t exist then, no amount of money could have bought it. Present-day she did eventually consume a profoundly expensive good, which sold for hundreds of thousands of dollars. 3/

But in welfare terms, was she better or worse off than her 20-years ago-prior doppelgänger? 4/

Ex post, no, she struggled more and still died. But even ex ante. She had some probability of being saved that her predecessor would not have had. But the ex ante probability of failing to receive the treatment, or of receiving it only with ruinous delay, would be high for many patients like her. 5/

Ex ante, the expected expenditure, and therefore her expected consumption level, would be much higher, but it’s not at all clear that her expected welfare is improved given high probability, welfare negative outcomes. 6/

When systems are misarranged, people can both be wealthier in real terms in the ways that we would measure that, and also worse off in welfare terms despite all that. /fin