In his 1970 book The Gift Society, British social scientist Richard Titmuss criticized the American practice of paying for donations of whole blood. He contrasted Great Britain, which relied on donors, and which seemingly faced far fewer shortages. Paying for other-regarding behavior, the book suggested, could crowd out more purely altruistic acts. Titmuss’s work attracted a lot of attention on this side of the Atlantic, including a book review by Kenneth Arrow, and the U.S. soon shifted to the British model.
But there has been some movement in the last decade back towards monetary incentives for donations of human body parts. More than a dozen U.S. states now offer tax incentives for living organ donors, although the incentives allow deductions only for the actual costs of lost wages and medical care.
Past studies of these state laws have mostly found, much as Titmuss suggested, that they are ineffective or even reduce donations. (Though one study found that tax incentives do encourage donations of bone marrow, which is among the least costly and invasive of the possible procedures).
These studies have all followed the basic difference-in-differences design common to empirical studies of U.S. legal change. The researchers draw a trend line for “treatment” states, and a trend line for “control” states that didn’t enact an incentive, and look for statistically significant differences in the slopes of the lines post-treatment. A key assumption, of course, is that any observed difference is the result of the treatment; that is, one assumes the lines would have been parallel if not for the legal change.
What happens if there are good reasons to suspect that assumption is false? For example, what if states that enacted organ-donor incentives were states where attention to the need for organ donations had recently surged? After enactment of the law, those states might regress to the mean, particularly if advocates relaxed, assuming that the new law would do some of their work for them. Comparing trend lines would then give our researcher a “spurious” negative correlation: it would look like enacting the law reduced donations, when actually the drop was the result of other forces. If those other forces were powerful enough, they might conceal any positive contribution of the law itself.
In other words, even though prior literature has reached a consensus that tax incentives for organ donation don’t work, or even reduce voluntary donations, we should be very skeptical of those papers. In a new project, my co-author Firat Bilgel and I try to study the impact of these laws in a way that doesn’t depend on the assumption of parallel state trends. Tune in next time for more details!