The behavioural economics of exercise habits

Jeremy D Goldhaber-Fiebert, Alan M Garber 22 February 2011

a

A

In the US, obesity – and the chronic diseases it can cause – is putting health policy under enormous strain (Flegal et al. 2010). One recent study finds that between 1993 and 2008 obesity was a greater threat to the health of Americans than smoking (Jia and Lubetkin 2010). But it is not just the US. The rise in hypertension, type 2 diabetes, and cardiovascular disease may have been most pronounced in the developed world, but many rapidly developing countries including India and China are now following this disturbing trend (Sugerman et al. 2003 and Mathers et al. 2006).

Increasingly sedentary lifestyles of many adults partially explain the increase in obesity over recent years. The typical American spends many hours in front of a computer or TV screen, failing to exercise – all in spite of the health benefits of doing so. Exercise is known to significantly improve cardiovascular health, blood pressure, and cholesterol and to reduce the risk of developing diseases such as diabetes and osteoporosis (Myers 2003).

For most people, failure to exercise is unlikely to be due to ignorance or lack of interest. It is more likely the result of difficulties in translating initial motivation to exercise into sustained behavioural change. In other words, it is hard to form an exercise habit. When asked, many American adults claim that they want to be healthier, to lose weight, and to be more active (Williamson et al. 1992).

A key problem is the difficulty of translating current motivation into long-term behavioural change. While for many people the prospect of exercising for three months and improving their health may seem attractive today, the daily decision to actually do the exercise during that period is a much more difficult one.

What approach do we take and what basic insights do we hope to gain?

In recent research together with Erik Blumenkranz (Goldhaber-Fiebert et al. 2010), we take a behavioural economic approach to the problem of helping individuals form exercise habits. We start from the premise that the longer someone actually engages in an activity the more likely that it will continue to be maintained as a habit. That is, their default will shift to doing the new action and their preferences may shift such that the default is more valuable.

Our challenge then is to design tools that enable individuals to engage in exercise long enough for it to become habitual.

We believe that even for the highly motivated, the daily challenges and demands of life often push them away from exercise before it becomes a habit. In behavioural economics, the problem of wanting to do something in the future but failing to do so when the future becomes the present is called “dynamic inconsistency” (Laibson 1997). A typical approach to dynamic inconsistency is to make sure that when an individual faces a choice in the future, the incentives he faces then cause him to act in accord with his original desires.

Consider a situation where I want to eat more apples and less cake to improve my heath, but although I say I will eat an apple tomorrow instead of cake, when tomorrow comes, I am sorely tempted and will eat the cake. If I could create a situation that increased the cost of eating the cake, I will be more likely to stick with my original decision and eat the apple. I could do this by signing a contract with myself now that states that if I do not do what I commit to now, I will pay a monetary penalty. This type of contracting with one’s self to sustain current actions in the future is a called a “commitment contract”.

Our interest, then, is two-fold:

  • We are interested in the idea of using exercise commitment contracts to promote exercise habit formation.
  • We also want to learn how long – regardless of the use of contracts or other means – must exercise be sustained before it becomes habitual.

To examine this second – more fundamental – question, we randomly have people exercise for different numbers or weeks and then watch how well they form and sustain exercise habits. This is challenging to do in practice because the people who choose to commit to more weeks in their initial exercise programme may differ from those who commit to shorter periods of exercise.

To address this issue, we use another behavioural economic concept – our susceptibility to default values, anchoring and “nudges” (Thaler and Sunstein 2009). For example, if I am first shown the number “4” several times and then shown three numbers “2”, “4” and “18”, in many cases I will be more likely to choose “4”. In our case, we use “nudges” for the default suggested length of the exercise commitment contract that individuals see when they are creating the contract. To the extent that this shifts their chosen contract duration and to the extent that they exercise more because of this randomisation, we can study the effects of random amounts of exercise on long-term exercise habit formation.

What has been done previously and what does our work add?

Savings and savings habits for retirement

Nudges have been used successfully to increase retirement savings. An increasing body of literature demonstrates that when default enrolment in retirement savings is used or when the default rate of savings is higher, more people save more money for retirement (Choi et al. 2004 and Benartzi and Thaler 2004). This is strong evidence for the power of “opt-in” versus “opt-out” programmes for two reasons:

  • Individuals have strong incentives to think about their money especially as it relates to decisions about current consumption and future savings.
  • It would be very easy for individuals to change away from the default value if they desired to.

Yet, they often do not. In this work, once the “nudge” has worked, no further action on the part of the individual is required to achieve the beneficial effect of higher savings rates.

Lessons from anti-smoking commitment contracts

Commitment contracts have been used successfully to help people quit smoking. Karlan and colleagues have shown that the use of commitment contracts leads more smokers to quit (Gine et al. 2010). In an important sense, quitting is the opposite of the problem we seek to address. Individuals must refrain from an action – smoking – long enough for the power of their previous physical addiction to wane. Repeatedly refraining from smoking until a new habit, which is the lack of an action, is formed is the name of the game.

Previous studies on exercise

The use of financial incentives to increase exercise beyond the period in which incentives have been given has also been shown to be effective by Gneezi and colleagues (Charness and Gneezi 2009). Clearly, incentives are important motivators for continued exercise. However, these studies were conducted primarily among younger adults – college students who, because of their age, tend to recover quickly from physical activity and have lower demands on their time than when they are older. Thus they may face lower obstacles to exercise than working-age adults. Further, adapting financial incentives for use with large groups of individuals – taking it to scale at the population level – while effective, may be an expensive proposition and potentially inefficient, since a portion of individuals who would have exercised even without the incentive would take it up (Kane et al. 2004 and Sutherland and Christianson 2008).

Our research focus

Our research aims to unite these types of findings (i.e., from retirement savings, smoking, financial incentives to exercise), focusing on middle-aged and older adults as individuals who could derive substantial health benefits from increasing their levels of physical activity.

What have we found and what do we hope to find in future?

Our first study shows that default contract length suggestions can nudge people into longer duration exercise commitment contracts.

  • Among 619 individuals aged 18-69 years who were randomly shown a default contract duration of “8 weeks”, “12 weeks”, or “16 weeks”, those shown a longer default duration were more likely to choose a contract of a longer duration – with a large proportion simply choosing whatever default value they were shown.
  • The results thus far show that individuals who were nudged into longer contracts did not commit to less exercise per week, nor did they choose smaller financial penalties for not fulfilling their longer contracts.

In subsequent follow-up work that we are now conducting in a larger sample of over 1,500 individuals, preliminary findings suggest the nudge effect is strong and consistent. Furthermore, we are also finding in our preliminary analyses that that individuals "nudged" into longer contracts through longer default suggested values successfully completed at least the same number of weeks of exercise and, importantly, completed more total exercise sessions during their contracts.

A number of outstanding questions remain.

  • First, in our current studies we rely upon self-reported exercise behaviour.

Future studies should add objective measures of physical activity such as the use of pedometers or other monitoring devices.

  • Second, our current studies focus on people who, although perhaps having difficulty exercising on their own, have found an online exercise commitment contracting website (StickK.com).

Future studies should be conducted in the context of employer wellness programmes or other populations more representative of the wide range of individuals whose health could improve through adopting long-term exercise habits.

  • Third, although we seek to promote long-term exercise habits by increasing the duration of initial periods of exercise – like those that result from exercise commitment contracts – the ultimate goal is to improve health outcomes.

If individuals were to increase their consumption of sugary beverages as they exercised more, believing that they can spend a little bit of their “extra healthfulness”, they might not improve their health overall. Future studies should not only measure increases in exercise in the post-contract “exercise habit maintenance phase” but should also include measures of health such as weight, height, pulse rate, blood pressure, cholesterol, and blood sugar.

Conclusions

Understanding the path from a sedentary lifestyle to a long-term exercise habit offers great promise for designing better prevention interventions aimed at serious chronic disease risks that now threaten large segments of the world’s population. Achieving effectiveness at a reasonable cost is imperative given the large numbers of people that must be reached. Behavioural economic approaches to this problem are promising and justify further research.

References

Benartzi S, RH Thaler (2004), “Save more tomorrow: using behavioural economics to increase employee saving”, Journal of Political Economy. 112(1):S164-S187.

Choi JJ, D Laibson, B Madrian, A Metrick (2004), “For better or for worse: default effects and 401(k) savings behaviour”, in Perspectives in the Economics of Aging (ed. David Wise), University of Chicago Press.

Charness G, Gneezi U (2009), “Incentives to exercise”, Econometrica, 77(3):909-931.

Flegal KM, MD Carroll, CL Ogden, LR Curtin (2010), “Prevalence and Trends in Obesity Among US Adults, 1999-2008”, Journal of the American Medical Association, 303(3):235-241.

Goldhaber-Fiebert JD, E Blumenkranz, AM Garber (2010), “Committing to Exercise: Contract Design for Virtuous Habit Formation”, NBER Working Paper 16624.

Gine X, D Karlan, J Zinman (2010), “Put Your Money Where Your Butt Is: A Commitment Contract for Smoking Cessation”, American Economic Journal: Applied Economics, 2(4):213-235.

Jia and Lubetkin (2010), “Trends in Quality-Adjusted Life-Years Lost Contributed by Smoking and Obesity” American Journal of Preventive Medicine. 2010; 38(2):138-140.

Kane RL, PE Johnson, RJ Town, M Butler (2004), “A structured review of the effect of economic incentives on consumers’ preventive behaviour”, American Journal of Preventive Medicine, 27(4):327-352.

Laibson D (1997), “Golden eggs and hyperbolic discounting”, Quarterly Journal of Economics, 112(2):443-477.

Mathers CD, Loncar D, 2006 Projections of Global Mortality and Burden of Disease from 2002 to 2030. PLoS Med,3(11): e442.

Myers, J (2003), “Exercise and Cardiovascular Health”, Circulation. 107:e2-e5.

Sugerman HJ, LG Wolfe, DA Sica, JN Clore (2003), “Diabetes and hypertension in severe obesity and effects of gastric bypass-induced weight loss”, Annals of Surgery, 237(6):751-758.

Sutherland K, JB Christianson (2008), “Impact of Targeted Financial Incentives on Personal Health Behaviour: A Review of the Literature”, Medical Care Research and Review. 65(6):36S-78S.

Thaler RH, CR Sunstein (2009), Nudge: improving decisions about health, wealth, and happiness. Penguin Books.

Williamson DF, MK Serdula, RF Anda, A Levy, T Byers (1992), “Weight loss attempts in adults: goals, duration, and rate of weight loss”, American Journal of Public Health, 82(9):1251-1257.

a

A

Topics:  Frontiers of economic research Health economics

Tags:  health, Behavioural economics, sport, exercise habits

Henry J. Kaiser Jr. Professor at Stanford University

Assistant Professor of Medicine at Stanford University