Tuesday, 28 February 2012

Behavioural blackboards

Behavioural economics usually leaves a bad taste in my mouth. Too much of it just feels like market failure theory as it was done in the 50s and 60s. Back then, you'd find papers with 20 pages of math proofs about how markets could never yield theoretically ideal outcomes, followed by a short paragraph talking about how government should therefore institute some policy to fix things; comparative institutional analyses weighing both the failures of markets and those of politics were rare. Substitute lab experiments for the math proofs and you've got rather too much of the behavioural economics. Niclas Berggren found that while a fifth of behavioural papers proposed paternalistic interventions, less than five percent worried about whether the behavioural anomalies giving rise to the policy interventions applied also to those who'd run the system. I'd touched on some of these issues in my talk at the CIS last week.

And, just as folks used to work backwards from the desired policy intervention to the market failure theory that could justify it, they can now work backwards from the desired outcome to the behavioural economics rationale. So if you don't like people getting compensation for organ donation or blood donation, just say that it'll interfere with altruistic donation and cite some experimental evidence on warm glow.

Mario Macis finds that, contrary to the behavioural blackboard, real world field experiments giving people real money for blood donations increases donations
As an economist, I wondered: Can we use economic incentives to stimulate blood donations? Economic intuition would suggest that the answer is yes, but theories in psychology and behavioral economics indicate that incentives might actually backfire if they conflict with people's intrinsic motivations to perform altruistic acts.

Together with Nicola Lacetera and Robert Slonim - my two research colleagues - I wanted to find a more definitive answer. We conducted a randomized controlled trial involving nearly 100,000 individuals, in partnership with the American Red Cross in Northern Ohio, to study the effect of economic incentives on blood donations.

In our study, we found that an advertised offer of a $5 gift card increased the likelihood of giving blood by 26 percent. A $10 gift card produced a 52 percent rise. And a $15 card caused an uptick of 72 percent. The offer of gift cards even caused people to motivate others to donate, including people who previously had never given blood. The incentives also induced people to switch the locations and dates of their usual blood donations to sites where the rewards would be available. Also, there was no effect on the share of individuals who presented at the drives but were ineligible to donate for medical or other reasons.

We concluded that economic incentives significantly increase donations from the public and can be used to increase donations in areas and at times of the year when blood supplies are particularly low.
Here's the paper (ungated). Some of the effect is due to displacement - some donors move to areas providing the incentives. But the result holds: supply curves slope upwards.

Moving to donor payment for blood might still not pass cost-benefit analysis: because you also have to pay for inframarginal units, the cost of the whole scheme in terms of new units donated wound up being $42 per unit. They reckon this easily passes cost-benefit given some estimates of the value of blood product collected.

Alas, under public health systems, the fiscal costs would fall on the government while the benefits would accrue to folks then able to get needed transfusions.

Update: Lynne Kiesling agrees on behavioural econ...


  1. Short question: if you give me $5 so I give you my blood is that still considered a donation?

    1. I'd expect that would be illegal. System also doesn't like directed donation.

  2. In fairness to colleagues who do behavioural economics, according to the data used by Berggren just under 80% make no policy recommendation. At the minimum this suggests the vast majority of articles that meet Berggren's restricted definition of making policy recommendations had reviewers that do not believe such thoughts pass scrutiny. My personal experience is that most experimenters are too 'academic' to think about policy, and those that do tend to be skeptical that any human action could be fully rational (the influence of literature on cognitive biases).

    That said, in a modern version of "madmen in authority, who hear voices in the air" the sillier bits of NZ's Ministry of Economic Development appear to believe behavioural economics is proof that Muldoon got some things right...

    David Smith

    1. If a fifth made recommendations, then four-fifths didn't; agreed. Of those making policy recommendations, only a quarter considered comparative institutional analysis.

      Worrying that anyone could pine for Muldoon.....

  3. just to confirm from commercial work that payment$ definitely do stimulate volumes at blood harvesting clinics in the usa. it can all get a bit dodgy though - dark times, desperate measures as Bulletproof might say.