Is Obesity Rational?
The rise of obesity in the U.S. is a major public health concern. Nearly one in three adults were obese in the early 2000s, up from one in seven a quarter-century earlier. If current trends continue, the obesity rate will rise to over 40 percent by 2020. Obesity is linked to diseases such as diabetes, hypertension, and asthma and to premature mortality - if left unchecked, the rise in obesity threatens to reverse historical gains in life expectancy. The rise in obesity is also an important contributor to the growth of health care costs in the U.S.
To understand the rise in obesity, it is important to understand how individuals make decisions about eating and other behaviors that affect body weight. In Understanding Overeating and Obesity (NBER Working Paper 16149), researcher Christopher Ruhm lays out two models of eating and weight regulation and examines empirical evidence to test which model is more consistent with people's behavior.
Ruhm begins with the rational economic model. In this model, individuals trade off the utility (or happiness) from eating food today against the expense and disutility of future weight gain. Individuals make the best decisions they can given the information they have and the constraints on their income and time. Under specified conditions, decisions are optimal (utility-maximizing) at the time they are made, even if the arrival of new information makes the decision look like a mistake in retrospect. In this economic model, factors such as falling food prices are the most likely explanation for the rise in obesity.
Of course, eating decisions may reflect biology as well as economics. Humans have been genetically programmed over millions of years to eat, with the primary goal being to obtain enough calories for survival. In an environment where food is cheap and readily available, biological programming may lead to overeating.
Ruhm formalizes this intuition in a "dual decision" model. The key insight of the model is that eating decisions are influenced by two parts of the brain, the "affective" system and the "deliberative" system. The affective system coordinates sensory inputs to generate emotional states like anger or happiness. This system responds to cues and stimuli - for example, the presence of food may generate endorphin and dopamine responses. The deliberative system incorporates higher cognitive processes, such as abstract thinking and planning, which account for long-term consequences of actions. In the dual decision model, "eating behaviors reflect the combined influence of a utility-maximizing deliberative system and an affective system that responds quickly and impulsively to external stimuli, without accounting for the long-term consequences."
The dual decision model has a number of implications for body weight and eating behavior. First, this model predicts that many people will be heavier than their utility-maximizing weight and will attempt to lose weight. In the rational model, by contrast, weight gain is utility maximizing and will rarely be accompanied by weight loss attempts.
Second, the dual decision model suggests that people with poor self-control or strong affective systems will be particularly susceptible to making eating mistakes. Since these people tend to have high body-mass index (or BMI, a measure of body weight calibrated by height), behaviors such as weight loss attempts will be concentrated among this group. In the rational model, there is no expected correlation with BMI.
Third, in the dual decision model there is more potential for "food engineering" (the strategic manipulation of food characteristics such as fat, sugar, and salt content to increase food consumption) to affect weight. In the rational model, food engineering can change an individual's optimal weight by affecting the pleasure obtained from eating, but in the dual decision model it can activate the affective system as well. Affective system responses will be stronger among those with high BMI, so overeating and the consumption of engineered foods will increase more over time among this group.
Ruhm uses data from the National Health and Nutrition Examination Surveys and the Behavioral Risk Factor Surveillance System to explore which model better explains trends in body weight, weight loss attempts, overeating, and consumption of engineered foods.
He has several key findings. First, body weight has risen rapidly over time, particularly in the right tail of the distribution - the BMI of men at the 90th and 95th percentile rose by 4.5 and 5.5 points over the past 25 years, versus 2.1 points at the median. Explanations related to the rational model, such as falling food prices, appear insufficient to explain these trends, since food prices fell only during the early part of the period while the rise in BMI continued throughout the period.
Second, the desire to lose weight strengthens with BMI, is pervasive among overweight and obese individuals, and has become more common over time. High-BMI individuals also experience greater fluctuations in their weight, which is consistent with their having periods of heavy eating combined with repeated efforts to lose weight.
Third, obese individuals consume disproportionate amounts of salt and fat, and the relationship between BMI and this consumption of salt and fat has strengthened over time. This is consistent with a greater role for food engineering as a cause of obesity over time.
In summary, the available evidence on body weight, weight loss, and food consumption is consistent with eating "mistakes" related to the dual decision model and is hard to reconcile with the standard utility-maximizing model.
These findings have important implications for policy. If the affective system plays an important role in eating decisions, this will tend to reduce the effectiveness of anti-obesity policies that rely on rational decision-making, such as taxing foods with high fat, sugar, or salt content or providing more information on the content of foods. A strong affective system may also point to the usefulness of policies advocated by behavioral economists, such as altering the location of food in a cafeteria line or reducing portion size in order to manipulate food-related stimuli.
Ruhm concludes "mistakes are a central feature of dual decision-making, implying a wider potential role for policy." However, he cautions "the specific interventions will often be complicated and, if poorly implemented, will reduce rather than increase utility. The general reluctance of economists to engage in policy activism therefore retains merit."