VoxEU Column Health Economics

Are you what you eat?

Smoking, heavy drinking, and being overweight are known causes of disease. This column presents experimental evidence to try and understand why people ignore this advice. It compares lifestyle choices with people’s attitudes to risk and their patience, finding that while people with an unhealthy lifestyle are no more risk-loving than other people, they are more impatient.

Smoking, heavy drinking, and being overweight all increase the chance of contracting a number of chronic health diseases – including cancer, cardiovascular diseases and diabetes (e.g. JAMA 2008). These diseases, in turn, are the major drivers of healthcare expenditure in developed countries.

Despite wide evidence associating these behaviours with disease, people often engage in unhealthy behaviours.

One possible explanation is that health behaviour may be intimately linked to deeply rooted individual characteristics, such as preferences, attitudes, and psychological traits.

Recent insights from economics and psychology, for instance, suggest that risky health behaviour can be explained in part by the way people perceive risks and evaluate the benefits of events that occur in the future (see also DeSimone 2010 and Duflo 2009).

Different people have different attitudes towards risk and perceive risky situations in dissimilar ways – in other words, their risk preferences are different. What makes risk in health distinctive from risk in other contexts, however, is the time delay that occurs between the actual behaviour, such as smoking, and its unwanted consequences, such as a cancer that forms much later in life. This makes the notion of time and how people perceive gains and losses in the future (time preferences) important factors in understanding health behaviour. Thus, among other factors, individual preferences for time and risk seem two central aspects that require careful analysis in the study of individual behaviour in relation to health.

New insights from experimental data

In an experimental project funded by the Alan Williams Fellowship from the Centre for Health Economics at York, we have analysed the relationship between individual perceptions of risk and time, and health habits (Galizzi and Miraldo 2010). We invited a sample of 120 students and members of staff at the University of York to answer one questionnaire and two experimental tests. The questionnaire focused on individuals’ eating habits and dietary balance, smoking and drinking habits, as well as sport or physical exercise. The collected data allowed us to construct a number of detailed indicators of subjects’ personal health behaviour and life style.

One major advancement of our study is that individual nutritional habits are summarised not only by using the Body Mass Index (which has often been criticised as a partial and inaccurate measure of nutritional habits), but also by a more detailed and precise measure: the Healthy Eating Index (HEI). HEI is a global indicator of the individual overall nutritional balance, constructed according to the guidelines of the panel of nutrition experts at the US Department of Agriculture, and reflects how close the actual nutritional intakes by a person are to her optimal nutritional balance. The latter is computed according to the individual characteristics and the daily energetic intake on different nutritional categories - such as carbohydrates, proteins, fruits and vegetables, oils and fats.

Jam today or jam tomorrow?

In our experiment subjects were then asked to answer two experimental tests to elicit their risk attitudes and time preferences. In the first, participants were asked to make a series of choices. In each of them, they had to choose between two lottery tickets with different payments and different winning probabilities. The payments and winning probabilities were chosen in a way that, according to individual risk preferences, one lottery ticket could be perceived as riskier than the other. By observing the individual choices for each pair of lotteries, our data allows the estimation of the underlying risk preferences of each subject, and to categorise the participants in terms of their estimated subjective risk aversion.

At the end of our experiment, participants were paid by cash according to the outcomes of the lottery tickets they chose in the test. The real monetary rewards paid to participants guarantee that the choices observed in our experiment genuinely reflect the true risk preferences of the subjects, a link that cannot be taken as granted when the measures are based on survey questions on hypothetical situations.

The second test had a similar structure. Participants were asked to choose between two options: in one option subjects would be given an amount of money immediately, paid cash at the end of the experiment, while in the other option subjects were paid, by cheque, a higher amount of money but only after some days. The amounts and delay of payments were chosen in a way such that, according to individual time preferences, one option could be perceived as more attractive than the other. By observing the individual choices at each pair of option, the experimental data allows to infer the latent time preferences of each participant, and to categorise the subjects based on their impatience and the way they traded-off immediate and future gratifications. In particular, we estimated a set of subjective traits that characterise hyperbolic and present-biased time preferences.

Previous evidence from experimental studies, in fact, has shown not only that we are typically impatient, in the sense that we prefer immediate rewards even at the cost of forgoing higher rewards in the future; but also that our impatience is typically biased towards gratifications that are closer to the present. For instance, when asked to choose between £100 today or £120 tomorrow, most of us would probably prefer to get £100 today. However, when also asked to choose £100 tomorrow or £120 in two days time, the same subjects would tend to prefer £120 the day after tomorrow. The evidence on this subjective perception of time (often mentioned as hyperbolic discounting) has attracted an increasing attention from social scientists, as it may challenge the often postulated idea of human beings as dynamically consistent decision-making agents.

Using the data collected under the questionnaire and experimental tests we explicitly looked at the relationship between individual time and risk preferences and lifestyle. In particular we analysed whether participants that had different eating and health habits made different choices in the risk and time preferences tests.

We find some evidence that subjects with different health habits also perceived risks and future rewards in a significantly different way. In particular, and contrary to what is commonly assumed by a vast literature in health studies, smokers did not appear to be significantly more risk-loving, while subjects with healthier nutritional balance and lower consumption of alcohol seemed to be more averse to risky situations. Also, smokers and individuals with unhealthy diets were typically more impatient, in the sense that they perceived immediate rewards and gratifications as more important than future costs, and, in particular, were found to have hyperbolic or present-biased time preferences.


If confirmed by further research on larger-scale samples, the evidence from our experimental project seems to suggest that there may be room for dealing with the diffusion of risky health habits, such as smoking and excessive eating, by looking at the subjective attitudes related to this compulsive behaviour, and, in particular, at the way subjects with different lifestyle perceive health gains and risks in the future.

By contributing to a better understanding of the underlying preferences of subjects engaging in behaviour detrimental to their health, our experimental analysis can contribute towards the design of public health policies and informational campaigns targeted at the reduction of obesity incidence, smoking and binge drinking.


Galizzi, Matteo M and Marisa Miraldo (2010), “Are you what you eat? Experimental evidence on health habits and preferences for risk”, Working paper funded by the Alan Williams Fellowship from the Centre for Health Economics at York.

JAMA (2008), “Smoking and the Heart”, JAMA Patient Page, derived from JAMA, 299(17), 7 May.

DeSimone, Jeffrey (2010), “Binge drinking and risky sex among college students”, VoxEU.org, 30 May.

Duflo, Esther (2009), “AIDS prevention: Abstinence vs. risk reduction”, VoxEU.org, 20 April.

735 Reads