Incentives for Nothing or Why Behavioral Experiments are Needed? Uri Gneezy talks in Barcelona

The room was full of people from Banco Sabadell, faculty members and PhD students from UPF, UAB, and UB interested in the research in the field of behavioral economics and judgment & decision-making. Barcelona’s prominent scholars in the field have also attended this event including Robin Hogarth (who chaired the talk by the way), Rosemarie Nagel, Gaël Le Mens, Antoni Bosch-Domènech among many others.

‍

It was an interesting talk for me for several reasons. First of all, Uri does a great job in popularizing behavioral and experimental research among business audience. It is important that business audience is driven no only by realizing the costs of mistakes (e.g. Uri’s example of Netflix rolling out new subscription plans in USA few years back), but also by realizing potential benefits (read: profits) from field testing their products and services with a pool of consumers using behavioral research methods. Again, stealing example from Uri’s talk:

Most of us are careful when making donations and selecting the organization to donate to. We are concerned whether our money will reach those who needed them and not used for “overheads” costs of charity organization itself. So, Uri and his colleagues have designed a field experiments where they offered different groups of people to donate money for some charity activity (read more about Randomized Controlled Trial (RCT) format on Wikipedia). Each of those different groups including a control group received information about possibility to donate but in a different format. While one group received a “standard” or classical donation request (this is our control group), other groups had donation requests where, for example, the donation request suggested that there is some organization that would match every dollar that a person donates. So if a person donates $5 than this organization matches donation of the this person and also donates $5.Playing with different formats donation formats, Uri and his team were able to identify the best donation request format. So, the best donation request format was when this request clearly stated that there is some organization which would cover overhead costs if a person donate let’s say 10 dollars thus making this person confident that money will go only for donation purpose.

“Incentives for Nothing” was another theme of Uri’s talk. Do incentive schemes at your organization work? How do you know? And that’s was the point that not many organizations are actually test whether the proposed incentives do make a change in people behavior. In yet another example:

Uri talked about “health plan” among employees of one financial institution in US. Concerned (? 🙂 ) about health of its workers, institution wanted to test existing and new incentives that would nudge employees to attend a gym lets say at least 5 times a month. Organization was already paying scaling up incentives for each time their employees went to gym. For example, if employee went once – he or she got $2. If he or she went twice than $5 and so on. If employee went 5 times than the payout would be $50. These are just made up numbers, but the idea should be clear now.Interesting part is that organization was already spending on this (on employees, contract with gym etc.) $100 million a year and getting little results.So, behavioral experiment was design to test new possible schemes that would improve participation rate. Designing different formats for different groups participating in the field experiment, the outcome of this research was unpredictable. When offering employees to pay $50 only if they attend 5 times and not less – nothing has changed in terms of participation rate. The “gut feeling” would be that at least people who attended 4 times a month would attempt to make it 5, but not! So, after this financial organization implemented this change in the incentive scheme ($50 only for 5 times a month) nothing has changes from the health perspective. However, now this organization was spending only half or $50 million a year.The learning point is that behavioral research might not give the right answer (or if it exists), but field experiments may help your organization to learn more about the efficiency of your existing policies or give insights on consumer behavior.

‍

In the following ‘cava’ reception, I have asked Uri whether he thinks that behavioral tools that lead to behavioral changes (or nudges) do actually work in the long run. Behavioral research and number of experiments has shown that nudges work (for main reference, I suggest you take a look at the book on nudges by Richard Thaler and Cass Sunstein here: http://en.wikipedia.org/wiki/Nudge_(book)), but what happens when people are repeatedly exposed to nudges? Uri suggested that indeed the effect of individual nudges like Tobacco warning labels may be arguable or not so clear, however step-by-step these and other nudges do reduce number of smokers. Whether you believe in libertarian paternalism or not, it is another big question.

‍

UPDATE: Check the highlights of the talk  in a 5 minutes video here

What do you think? Offer your thoughts or links to existing research on the topic in the comments section

‍

Also check some of our behavioral experiments that we did together with our partners for European Commission on Tobacco warning labels, Testing CO2/Car labelling options and Online gambling nudges.

• Study on online gambling and adequate measures for the protection of consumers of gambling services
• Testing CO2/Car labelling options and consumer information
• Study on tobacco labelling and packaging (coming soon!)

‍