Why nudges fail and other puzzles: insights from research on commitment devices

manu savaniManu Savani 

Having just read the new special issue and accompanying blog series published by Policy & Politics entitled Beyond nudge: advancing the state-of-the-art of behavioural public policy and administration, I was inspired to respond to some of the arguments mooted.

The question of why we find behaviour change resolutions difficult to stick to has long been the subject of debate and research. It is familiar territory at this time of year as we contemplate new year’s resolutions. Knotty inter-temporal choices can be affected by present bias, where we focus on short-term gains rather than the long-term payoffs. Commitment devices – any voluntary strategy we use to influence our future decisions and achieve our goals – have shown promise in addressing present bias. These strategies can rely on financial stakes, as shown by the stickK approach, which reports having $51 million on the line across 527,000 individual commitments.

Commitment devices can also rely on reputational stakes, where the main motivation to stay on track arises from wanting to keep our word to others and ourselves. Unlike other nudges, they are transparent and self-driven, and are less open to critique for being manipulative or opaque. But so far they have been limited in their policy applications, partly because they have rarely proven effective beyond the short-term.

As the behavioural public policy agenda matures, academics are increasingly looking at null results in policy experiments and asking why nudges fail – themes that this special issue blog series has highlighted effectively in previous posts. Commitment devices too provide an important testing ground for this question. Why do commitment strategies sometimes fail? Why do we fail to stay committed to our commitment strategies?

Fatalism may be part of the answer, following the argument in Tom Entwistle’s recent article for Policy & Politics. Another explanation is that people might feel overloaded by behaviour change commitments. But commitment strategies could also fail if they are too mild, which may be a particular concern with reputational commitment devices that are seen as ‘softer’ than those involving money.

My research studied the effectiveness of commitment contracts in a community health initiative run by a local government agency in the UK. I wanted to test whether these milder commitment devices could encourage people to attend to their health goals over several weeks, a bit like a new year’s resolution. A field experiment randomly assigned members of a weekly weight loss programme to receive, alongside their usual group resources, a commitment contract. By signing it, participants were enhancing their commitment to the 11-week programme, pledging to see it through to its completion, and to actively pursue their 5% weight loss goal.

Like other interventions highlighted in this special issue series, my research reported some good news and some puzzling news. On the one hand, the commitment contract raised attendance by 8%-12% on average, with 77% of members who signed the contract seeing the course through to its final session compared to 69% in the comparison group. Further, the results suggested that some participants, who reported myopic and fatalist health attitudes at the baseline, benefitted more from having the contract. These findings on heterogeneity in commitment contract effectiveness require careful interpretation, but do support the concept that commitment devices are particularly useful for those who might otherwise struggle to achieve their health goals due to behavioural biases.

Health providers might welcome the boost in participation caused by the commitment contracts, as they strive to retain members over the full duration of the programme. However, these improvements in participation were not matched in weight loss performance. There was no significant difference between the groups based on signing the commitment contract. Why, if participants were participating more, did this not affect their final health outcomes? To answer this puzzle, we need more research on ‘moral license’ effects, which are relatively under-studied in relation to commitment devices and behavioural interventions more widely. Could it be that the commitment contracts encouraged people to be ‘good’ in one aspect of their health behaviours, but gave them license to be less ‘good’ in other aspects?

In practical terms, my study implies that commitment contracts might be better suited to near-term, discrete goals such as ‘I will attend class this week’, but are less able to affect a nexus of behaviours that are required for more sophisticated goals like ‘I will lose 5% body weight’. There may be scope for community health initiatives and local government programmes to co-design commitment contracts with their users, and reaffirm them frequently to maintain their salience and relevance to their users’ daily lives. There is much still to explore with commitment devices, and embracing unexpected findings is the first step to solving these puzzles.

If you enjoyed this blog post, you may also be interested to read:

Why nudge sometimes fails: fatalism and the problem of behaviour change

Advancing behavioural public policies: in pursuit of a more comprehensive concept

Beyond nudge: advancing the state-of-the-art of behavioural public policy and administration

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s