I recently read David Halpern’s Inside the Nudge Unit. I would recommend it to anyone who’s work touches on policy implementation, service delivery or any kind of behaviour change. Here are some snippets I took from it.
The context for Inside the Nudge Unit
In 2008, Richard Thaler (a University of Chicago economist) and Cass Sunstein (a Harvard Law School Professor) published Nudge. Nudge was the book that captured, and brought to the mainstream, the ideas of behavioural economics, libertarian paternalism and choice architecture (and in someways was the natural predecessor of Daniel Kahneman’s Thinking, Fast and Slow).
At the time, the NYTimes wrote:
“an engaging and insightful tour through the evidence that most human beings don’t make decisions in the way often characterized (some would say caricatured) in elementary economics textbooks, along with a rich array of suggestions for enabling many of us to make better choices, both for ourselves and for society”
“In the end, however, “Nudge” is somewhat thin on practical ideas for public policy that follow from the authors’ core insight.”
In 2010, UK Government established the Behavioural Insights Team (BIT) (a.k.a the “Nudge Unit”). Inside the Nudge Unit is a collection of stories and projects from Halpern’s time inside BIT.
The “EAST” Framework
After working through a few different behavioural insights frameworks, BIT settled on the “EAST” Framework. The four EAST principles being:
- Easy: if you want people to do something, make it easy – Recognise the the power of defaults. Reduce the ‘hassle factor’ or ‘friction costs’ of taking up a service. Make messages simple.
- Attractive: draw attention to, and make attractive, the thing you want people to do
- Social: show people that the desired behaviour is normal – Don’t inadvertently reinforce problematic behaviour. Use the power of networks. Encourage people to make a commitment to others.
- Timely: at different moments, the exact same offer can have drastically different effects – People will often put much more weight in immediate costs/benefits rather than later ones. Habits are much easier to build (from scratch) rather than break.
Nifty examples illustrating the principles
Making home insulation easy
Many governments have subsidies for house insulation. Interestingly, many people do not take advantage of these. BIT found that combining house insulation subsidies with a loft clearance service (which was not free), the take-up was three times higher.
Making motorcycle theft hard (by mistake)
When motorcycle helmets were made mandatory (West Germany, Texas, Britain, Netherlands), motorcycle thefts declined: Motorcycle thieves can’t/wont/don’t carry helmets.
Making driving under the speed limit attractive
In Sweden, drivers that stayed under the speed limit were recorded and entered into a lottery. A prize was given to a winner, this prize money came from a portion of the fines to those over the limit. Average speeds dropped from 32kmh to 25kmh.
Making direct debit payment attractive
Some Councils in London provided a lottery prize to those who switched to direct debit payments. The saving generated were 10x the cost of the prize.
Signally that littering is socially acceptable
After finding a flyer on their car, people are eight times more like to drop it on the ground if the carpark is already littered with other flyers.
Changing the wording on income tax letters to make relevant social norms explicit
By changing the wording on letters to people who were late to pay their income tax, BIT found by adding the line
“nine out of ten taxpayers pay on time”
they could increase payment rates by 4.5 percent. This could be raised to 16 percent by making the line even more relevant:
“most people in <your local area> pay their tax on time […and..] most people with a debt like yours have already paid”
Proposing timely changes to commuting habits
“Quite often policy makers are so focused on the ‘what’ — the intervention at hand — that they pay little attention to the critical question of ‘when’ (and ‘where’) .
For example, imagine you want to encourage more people to walk, cycle or take public transport to work. There are lots of ways you might do this, such as putting in cycle lanes (easy), emphasising that many others walk/cycle to work (social), or making public transport more comfortable and frequent (attractive). Some governments have even funded one-to-one transport advice for workers, intended to encourage more to leave their cars at home.
Unfortunately, many of these strategies are not very effective, although there is one big exception — among people who have just moved house [or changed employer -ML]. You can see why. If you’ve just moved house, your journey to work habits have yet to fully form, and you will be much more open to the suggestion of alternative options than if you have already been driving to work for the last five years.”
In another part of the book, Halpern points out that this is why new parents are so heavily marketed to: if your whole life is changing so much, changing your brand of butter or washing powder is a relatively tiny change.
Justice declines as the day progresses
Judges make positive judgements at the start at a rate of 65%, but this declines to ~0% positive by late morning. After lunch, they return to ~65% and then drop later in the day.
My takeaways from Inside the Nudge Unit
Baked into the establishment of BIT was a doomsday clause:
If, after 24 months, BIT hadn’t delivered at least a tenfold return on the cost of the unit, it would be disbanded.
Interestingly, this itself is a kind of commitment device. It is also uncharacteristically brave, not just in the public sector but anywhere, to counter the natural optimism bias of a new project.
Ultimately, BIT were so successful, that they were privatised and are now a social purpose company owned by the Cabinet Office.
I recently attended the ALGIM 2016 conference where a speaker claimed:
While I haven’t been able to verify the claim, the principle matches the meme de’jour:
the benefits of many small, well considered, incremental tweaks (or nudges) can add up to eclipse large complex projects
If you make a thousand $1 dollar bets, you only need a few good wins to break even. If you make one $1,000 bet, you only have one chance to make your $1,001 back.