Persistence or reversal? The micro-effects of time-varying financial penalties
Introduction
A now substantial body of evidence demonstrates that financial incentives can motivate behaviour change across a range of behaviours including smoking cessation (Notley et al., 2019), physical activity (Barte and Wendel-Vos, 2017), weight loss (Ananthapavan et al., 2018) and safe driving (e.g. Mazureck and van Hattem, 2006; Bolderdijk and Steg, 2011; Mortimer et al., 2018; Peer et al., 2020). However, the available evidence suggests that financial incentives don't always work as intended (e.g. Jeffery et al., 1993; Patel et al., 2016b; Volpp et al., 2006) and that the effects of financial incentives may vary depending upon a wide range of design parameters including: dollar-value, monetary versus non-monetary, temporary versus permanent, and constant versus time-varying (see, for example, Deci et al., 1999; Gneezy et al., 2011; Lacetera and Macis, 2010). The evidence also suggests that financial incentives may have different effects in different populations and settings (e.g. Gneezy et al., 2011; Notley et al., 2019; Paloyo et al., 2015). Despite this progress, significant gaps in our knowledge remain and it is not yet clear how best to structure (and adjust) financial incentives in each application. Further evidence is therefore required if we are to maximise the impact of financial incentives and avoid unintended consequences (Gneezy et al., 2011; Homonoff, 2018).
With regards to the permanence of incentives, a number of studies have investigated whether the short-term effects of temporary incentives persist after incentives are removed (e.g. Cahill and Perera, 2011; Charness and Gneezy, 2009; Halpern et al., 2015; Loewenstein et al., 2016; Volpp et al., 2008, 2009). This is an important question because many real-world incentive schemes are time-limited or time-varying by design. Such design choices may be motivated by the prohibitive costs of maintaining an indefinite scheme, or the belief that temporary incentives offer a better balance of costs and benefits.
This belief may be well-founded for temporary incentives that remain in place for long enough to push habit stocks beyond the threshold required for self-sustaining habits (Becker and Murphy, 1988; Lally et al., 2010).1 Habit formation may also be achievable in the short-run. Recent findings reported by Carrera et al. (2020) suggest that incentives that switch on and then off over time (‘sporadic on/off’ incentives) may induce temporary increases in habit stock and treatment effects that persist in the short-term but decay over time.2
Of particular relevance for the present study, we might expect a different effect from ‘sporadic on/off’ incentives in the presence of motivational crowding out (Frey and Jegen, 2001; Gneezy et al., 2011; Goswami and Urminsky, 2017).3 This is because it can be difficult to undo the reductions in intrinsic motivation that go hand-in-hand with motivational crowding out (Deci et al., 1999; Gneezy and Rustichini, 2000). Permanent reductions in intrinsic motivation raise the potential for behavioural ‘reversal’ rather than ‘persistence’ after incentives are removed (Gneezy and Rustichini, 2000) ,4 where the term ‘reversal’ refers to a deterioration in the target behaviour below pre-incentive or no incentive levels.5
The present study investigates the effects of habit persistence and behavioural reversals on the overall effectiveness of financial penalties for risky driving as they repeatedly switch ‘on’ and ‘off’ and as they weaken over time. We hypothesise that reversals during ‘off’ weeks can completely offset the positive effects of incentives during ‘on’ weeks, even when the incentive is only exhausted and switched off for a minority of participants. Reductions in incentive strength are hypothesised to further undermine the overall effectiveness of incentives.
Findings from the present study make a unique contribution to our understanding of the motivating effects of incentives and the circumstances under which these effects may be undermined. Our paper is among the very few to evaluate the micro-effects of temporal variation in incentives and its potential to generate behavioural reversals. We continuously tracked outcomes as the (loss-framed) incentive was repeatedly switched on and then off and as the strength of the incentive varied over time. In the only other study to evaluate the impact of sporadic on/off incentives, incentives during ‘on’ weeks were gain-framed and contingent on behaviours that are unlikely to bring intrinsic and extrinsic motivation into conflict (Carrera et al., 2020). Our focus on financial penalties also sets our study apart from the vast majority of field experiments and lab studies where gain-framed incentives have been employed (Jochelson, 2007), often with the aim of maximizing acceptability and/or participation (e.g. Halpern et al., 2015; Volpp and Galvin, 2014). For similar reasons, the majority of previous studies of loss-framed incentives have employed deposit-based schemes (e.g. Halpern et al., 2015, 2018; Volpp et al., 2008; White et al., 2020) rather than ‘fines’ or ‘penalties’ for non-compliance.6 Our findings therefore also make an important contribution to ongoing debate regarding the merits (or otherwise) of financial penalties.
Section snippets
Study design & hypotheses
This paper describes secondary analysis of data from a randomised field experiment comparing: placebo control, feedback on driving behaviour, and feedback on driving behaviour plus financial incentives for safer driving. The experiment was designed to identify the effect of feedback and financial incentives on three outcomes that are important road driving behaviours (speeding, hard acceleration and hard deceleration). In-vehicle telematics technology was used to collect data continuously on
Recruitment & study sample
Recruitment commenced in October 2017 and the final participant was randomised in late December 2018 (n = 232). Of the 232 participants who received their assigned treatment (ITT sample), 58 participants failed to complete at least one trip during either baseline or intervention periods and therefore recorded no outcome data in either baseline or intervention periods. We treated these 58 participants as lost to follow-up and took the remaining 174 participants as our final (ITT-exposed) study
Main findings
We investigate the effects of habit persistence and behavioural reversals on the overall effectiveness of sporadic on/off penalties for risky driving. Results suggest that incentives can influence driving behaviour while ‘switched on’ but may do more harm than good if they can also be ‘switched off’. Whereas feedback plus a ‘switched on’ incentive reduces the probability of speeding by an average of about 0.12 (p < 0.01), we find evidence for behavioural reversal (rather than persistence) when
Conclusion
While it is well-known that incentives can influence behaviour, relatively little is known about the micro-effects of incentives as they vary over time. This study demonstrates that variation in the presence and strength of financial penalties can undermine their effectiveness; potentially leaving affected participants (and society) worse off than if we had done nothing at all. We find evidence for reductions in the target behaviour relative to placebo control when incentives are
Ethics approval
University of Melbourne Human Research Ethics Committee (approval number 1646290.1).
Declaration of Competing Interest
None.
Acknowledgments
This work was supported by the Australian Research Council (LP150100680) and the National Health and Medical Research Council (APP1136250). The authors thank Katherine Scully and Mike Adams for assistance with administering the experimental protocol. We are also grateful to three anonymous referees for their thoughtful comments.
References (61)
- et al.
Effects of Pay-as-you-drive vehicle insurance on young drivers’ speed choice: results of a Dutch field experiment
Accid. Anal. Prevent.
(2011) - et al.
The relative caloric prices of healthy and unhealthy foods differ systematically across income levels and continents
J. Nutr.
(2019) Risk-taking attitudes and risky driving behaviour
Transp. Res. Part F
(2004)- et al.
The endowment effect and the reference state: evidence and manipulations
J. Econ. Behav. Organ.
(2009) - et al.
Do all material incentives for pro-social activities backfire? The response to cash and non-cash incentives for blood donations
J. Econ. Psychol.
(2010) - et al.
Pay as You Speed, ISA with incentives for not speeding: results and interpretation of speed data
Accid. Anal. Prevent.
(2012) - et al.
Habit formation in children: evidence from incentives for healthy eating
J. Health Econ.
(2016) - et al.
The effect of ‘smart’ financial incentives on driving behaviour of novice drivers
Accid. Anal. Prev.
(2018) - et al.
Who responds to financial incentives for weight loss? Evidence from a randomized controlled trial
Soc. Sci. Med.
(2015) - et al.
App-based feedback on safety to novice drivers: learning and monetary incentives
Transp. Res. Part F
(2020)