key: cord-1038518-6oqwjj93 authors: Sanders, Michael; Stockdale, Emma; Hume, Susannah; John, Peter title: Loss aversion fails to replicate in the coronavirus pandemic: Evidence from an online experiment date: 2020-07-24 journal: Econ Lett DOI: 10.1016/j.econlet.2020.109433 sha: 4420ba48ac1065a1653a8cc8fcee8ce727d44fcd doc_id: 1038518 cord_uid: 6oqwjj93 Loss aversion is a foundational bias and is a natural choice for interventions encouraging compliance during COVID-19. We compare the effectiveness of loss and gain messages and find no difference in the intention to comply with guidance or lockdown beliefs. First introduced in Kahneman and Tversky's (1979) paper on prospect theory, the concept of loss aversion is one of the oldest and most robust findings in behavioural science. Loss aversion refers to the fact that, when making judgements, prospective losses are felt more negatively than equivalent prospective gains, which are felt positively. Kahneman and Tversky (1979) found that individuals were more willing to take greater risks to avoid certain loss, but would avoid risk if the alternative was an equivalent certain gain. Small changes in the choice architecture such as the framing of loss have been shown to influence judgements in health (e.g. Farrell et al., 2001) , finance (e.g. Haigh & List, 2005) and environmental sustainability (e.g. Segev, et al, 2015) . Framing the current COVID-19 pandemic in terms of the potential loss of life from taking (or failing to take) some actions may inadvertently lead to loss aversion, which could motivate the very behaviours policymakers want to prevent (van Bavel et al, 2020) . This paper aims to explore the effect of gain and loss messaging on individuals' support for extended lockdowns and their intention to adhere to public health guidelines in this context. It is hoped that in doing so a greater understanding of how such nudges 1 differ in times of public crisis can be determined. Hameleers (2020) examined gain and loss framing on hypothetical COVID-19 policy preferences and emotional responses, finding that loss framed messages increase favour for a risk-seeking policy intervention, with a message framed as lives saved (gain) resulting in a more risk-averse approach. Prior research on loss aversion in the health domain has focused on relatively certain outcomes (e.g. quitting smoking will improve health). However, the measures taken by the UK government (and others) in the current pandemic are arguably less certain with an ongoing debate as to whether lockdown has substantially improved outcomes. Indeed, Rothman and Salovey (1997) suggest that frames vary in effectiveness depending on the certainty of the outcome. They posit that loss-framed messages are more effective when advocating for behaviour associated with greater uncertainty (e.g. STI testing) compared to advocating for certain outcomes (e.g. condom use) which is more effective with a gain-frame message. It may therefore mean that the uncertainty of the lockdown's effectiveness results in people paying more attention to negativeframe messages as they seek to minimise uncertain losses. To date, research on the effect of loss and gain messaging in a global health crisis was based on hypothetical interventions prior to the introduction of lockdowns (Hameleers, 2020) . Therefore, in this study, we explore how loss and gain messages about the novel coronavirus influence the desire to remain in, or reduce, the current lockdown policies currently in place in the UK and elsewhere. Given the robust findings of loss aversion, it is expected that: a message framed as a gain (saving lives) will result in a preference for curtailment of current lockdown compared to a message which is loss framed (loss of life), and a message framed as a loss will result in greater intentions to adhere to health J o u r n a l P r e -p r o o f guidelines compared to a gain-framed message. We registered these hypotheses with the Centre for Open Science 2 prior to commencing data collection. This research aims to test a hypothesis that many behavioural scientists, and the politicians they may advise, hold to be true: that they are able to 'nudge' behaviour, and in so doing improve outcomes, by their choice of a gain or loss frame. As will be seen, this need not be the case. We recruited 500 participants from Prolific (on 18 and 19 May 2020 J o u r n a l P r e -p r o o f Journal Pre-proof Participants were randomly allocated between two conditions: a gainframe and a loss-frame. In both scenarios, participants were shown estimates of the number of lives that could be saved through extension of the lockdown, or conversely, lost from ending it too soon: 3 Gain: As many as 100,000 people could be saved by a well-managed extension to the lockdown. Loss: As many as 100,000 people could die without a well-managed extension to the lockdown. They were then asked to make a series of judgements about when different elements of society should be opened up, and their intentions to comply with the government's guidelines. We collected participants' age and gender, as well as responses to a series of questions relating to the COVID-19 outbreak and how quickly the These questions were aggregated for analysis into a single measure of average days to guidelines being relaxed. Participants were asked also about how long (in weeks) they felt that pandemic response measures in general should continue, and a series of questions relating to their intention to comply with existing guidance. A full survey can be found in the appendices. We analyse our data using linear regressions, in which each of our outcomes is regressed on the loss-frame treatment, and on participants' gender and age. Our main results use aggregated scores from our compliance questions and average wait to relax restrictions questions; and the single-item measure of overall how long the lockdown should continue -see Table 1 . Differing from the established literature, we find no significant impacts of loss aversion on any of our three outcome measures. With two of these outcomes, compliance and average duration, participants seeing the loss frame are slightly (but not significantly) more in favour of faster easing, and less likely to comply. Given the size of our sample, and the historically strong effects within this literature, we do not consider our analyses to be underpowered. Journal Pre-proof the gain-framed message, but this difference was comparatively modest -fewer than 10 calendar days on average, and given the number of tests conducted, could well be spurious. Loss aversion did not affect peoples' preferences about lockdown nor their intention to adhere to public health guidelines as prospect theory suggests. One J o u r n a l P r e -p r o o f explanation for this departure is that this may be a similar scenario to that in Hallsworth et al., (2017) , which finds that in the case of tax liabilities, loss and gain frames are equally effective messages emphasising the public good. However, the different context of the present study may also contribute. The predicted death rate for COVID-19 is ubiquitous in the media and it is therefore likely that participants were familiar with anticipated death toll expected both with and without lockdown measures. Alternatively, the findings may depart from previous research including that of Hameleers (2020) et al., 2020) . Overall, the present study adds to the understanding of loss aversions in the context of a public health crisis and the boundaries within which gain and loss messaging may be effective. • How long do you think social distancing of 2 metres should continue until? Enter a date: • When do you think that office workers currently working from home should return to the workplace by? Enter a date: • When do you think the restrictions on international travel should be lifted? Enter a date: Participants were also asked the below question about the lockdown: • Overall, how long do you think the current social distancing and public health guidelines (i.e. not being allowed to leave the house for certain reasons, and shops and restaurants closed) will be in place for? The impact of the coronavirus lockdown on mental health: Evidence from the US. Human Capital and Economic Opportunity Working Group Reference-point formation and updating Confidence in the safety of blood for transfusion: the effect of message framing Do professional traders exhibit myopic loss aversion? An experimental analysis The behavioralist as tax collector: Using natural field experiments to enhance tax compliance Prospect Theory in Times of a Pandemic: The Effects of Gain versus Loss Framing on Policy Preferences and Emotional Responses During the 2020 Coronavirus Outbreak A model of reference-dependent preferences Lifting lockdown could cost 100,000 lives, says Neil Ferguson of Imperial When to release the lockdown: A wellbeing framework for analysing costs and benefits Comparing gains and losses Is loss-aversion magnitude-dependent? Measuring prospective affective judgments regarding gains and losses Shaping perceptions to motivate healthy behavior: the role of message framing The effects of gain versus loss message framing and point of reference on consumer responses to green advertising Prospect theory: An analysis of decision under risk Using social and behavioural science to support COVID-19 pandemic response Coronavirus disease (COVID-19) advice for the public Intention to adhere to public health guidelines: Questions asked to participants Participants were asked to answer the below questions regarding their intention to adhere to the public health guidelines • Working from home unless essential to do otherwise • Following hygiene precautions like washing your hands for 20 seconds • Social distancing from others apart from those in your household • Volunteering from the NHS • Staying home as much as possible • Stockpiling food and other household goods • Travelling to see friends and family • Wash hands more frequently and for longer than normal Inform others if I develop symptoms of COVID-19, no matter how mild • Spend time in public, sunbathing, sitting or picnicking • Disinfect frequently touched objects and surfaces • Go to a garden centre We acknowledge our funder King's College London. The grant is part of the King's Together COV-2 funding call. There are no competing interests. • How long do you think the government's furlough scheme, which pays 80% of staff's wages for those who cannot work due to COVID-19, should be continued for? Enter a date:• When do you think large gatherings (e.g. football matches, theatres) should begin? Enter a date:J o u r n a l P r e -p r o o f