Cambridge, Massachusetts, can claim to have one of the most outstanding collections of researchers working on behavioural economics in the world. It’s intent on getting even better, not least by marshalling resources across Harvard Business School (HBS), Kennedy School of Government (KSG), and drawing on MIT and Boston University too. It also has the impressive Harvard Decision Laboratory, a facility shared between 13 faculties, capable of running large volume trials, with multiple psychological and physiological measurements and access to a large pool of both student and general population subjects. In October 2012, David Halpern, Director of the Behavioural Insights Team, and Laura Haynes, Head of Research, visited Cambridge. Particular thanks go to Michael Luca of HBS for organising much of the visit. The trip revolved around talks at HBS, KSG and a joint talk given by Cass Sunstein and David Halpern. Cass is now back at Harvard, writing his latest (nudge relevant) book, and has just published his first post-government paper (on the case for personalised defaults rather than general defaults or unstructured choice)
Despite having just done an election all-nighter for the US election, Cass gave a passionate but measured overview of the impact that he thought behaviourally based interventions were having on policy, and the further potential impact they can have in the years to come. He gave as an example that deaths on US roads had just fallen to an all-time low – in absolute terms, not just relative – and argued that this was largely due to behaviour changes, such as driving habits, wearing of seat-belts and so on (note also the role played by disclosure around safety in changing consumer behaviour which in turn has put pressure on manufacturers to make cars safer: see Fung et al’s book on Full Disclosure, documenting how publication of roll-overs in safety tests have transformed the safety records of SUVs). Cass also used a very similar conceptual framework to ours, not least with a major emphasis on simplification, highlighting the executive order to reduce paperwork. It often takes several years for new research findings to filter out of academia, especially through the journals. This is problematic in a rapidly developing field. Talking to leading researchers directly shortcuts this process. Here are some of the headlines, organised around the short mnemonic ‘EAST’, that we sometimes use (Easy, Attractive. Social, Timely). E is for EASY
- Do opt-out pensions simply lead to less saving elsewhere? Raj Chetty (Economics) has used Danish data to show that when people are automatically opted into pension savings, it leads to very little reduction in other savings elsewhere. In contrast, widespread schemes that use tax incentives to encourage people to save more simply lead to shifts in modes of saving, and only affect the sophisticated minority that would have saved anyway. Chetty calculates that a $1 subsidy to save (at direct cost to the state) led to around 6c extra saving, while $1 in an automatic enrolment leads to around 80c extra saving.
- The impact of transparency measures depends on ease of interpretation. The effectiveness of transparency measures often hinges on how easy the disclosed information is to understand. Archun Fung et al show that while simple star ratings on how easily different SUVs roll-over on impact led to a rapid reduction in deaths (as manufactures scrambled to improve their ratings) the information available on water quality is too complex for consumers to understand and has had negligible effects on consumers and therefore companies.
- Gender bias in selection can be reduced by ‘bundling’ recruitment decisions, rather than doing them sequentially. In lab experiments, Iris Bohnet (KSG) shows that in one-off decisions, selectors will often choose a man with a worse score than a women (such as in a maths task) where the stereotype view is that men are better at that task (ie people think that men are better at maths, often despite the evidence). However, this effect was eliminated when selectors had to pick people for several such jobs at once, forcing them to look at a larger field of candidates. Bohnet argues that this may explain why men and women often do equally well in entry level selection (bundled) but men pull ahead in senior level appointments (one-off decisions – is X ready for promotion?), and recommends that firms therefore seek to bundle senior level selection decisions too.
- Commitment devices to boost saving. Savings products that restrict access to the cash are extremely popular, and effective, in developing countries. This work by Nava Ashraf (HBS) replicates similar work in OECD countries that people know they struggle to save, and would rather bind themselves in than expose themselves to temptation even at the cost of reduced flexibility.
- Simplification. We have often found that just simplifying a letter or communication into plain English can have big impacts. This work has attracted the attention of the prolific author, and Harvard professor, Stephen Pinker (Psychology) who is gathering such examples for possible future book (do send him some!)
In the next post, we’ll fill in the headlines under the second part of the mnemonic: Attractive.