May 17, 2016

Taking Paternalism Out of Nudge: The Case of Medication Nonadherence among Patients with Chronic Conditions

Contact us
To speak with a scholar or learn more on this topic, visit our contact page.

Over the past few years, the federal government and local governments have increasingly turned to “nudges” as solutions to many problems caused by behavioral biases. For example, policymakers have used nudges to influence consumer behavior via consumer finance regulation and antiobesity campaigns. Typically, these policies have been highly intrusive and paternalistic.

A new paper for the Mercatus Center at George Mason University examines how policymakers’ errors when diagnosing behavioral biases influence the effectiveness of paternalistic and nonpaternalistic behavioral policies. The paper demonstrates that, compared to paternalistic policies, nonpaternalistic nudges produce more positive outcomes for consumers.


Behavioral economics research shows that individuals’ decisions concerning their personal finances and health are often biased.

  • The impact of consumer bias can be significant. Studies have shown that more than half of Americans do not save enough for retirement, and the Centers for Disease Control and Prevention estimate that more than two-thirds of Americans are overweight or obese—conditions that may lead to substantial health problems.
  • Policymakers attempt to mitigate the impact. Seeking to combat biases in areas where mistakes are costly to consumers, policymakers have introduced an array of intrusive, paternalistic policies such as bans on large sodas and restrictions on payday lending.


Consumers have limited cognitive resources simply because mental resources are finite; using them for one activity means they may not be available for other activities. Scholars have identified several categories of limited mental resources.

Cognitive Capacity

People rely on two strategies to economize on mental resources and capacity:

  • One strategy is to use fast, intuitive thinking (heuristics). This may sometimes be efficient, but other times it gives rise to bias. For example, patients may opt for surgery merely because it is framed in terms of the chance of success rather than the chance of failure.
  • Another strategy is to use mental models to understand relationships and predict outcomes. For example, farmers in India often apply too much fertilizer to their crops because they incorrectly believe that green leafy growth signals overall crop health.


Individuals are assumed to value a payoff in the present more than the same payoff in the future. However, studies have found that people use time-inconsistent or “hyperbolic” discounting. Their valuation of payoffs differs depending on the time horizon.

For example, inertia or status quo bias may prevent employees from enrolling in company-provided retirement programs; consumers may also prefer to spend money when it is available rather than to save it.


Simple forgetfulness may be responsible for some suboptimal consumer choices. For example, many patients with chronic conditions fail to take long-term medication because they forget.

Additionally, complex decisions may lead to information overload. For example, one study found that a quarter of mortgage consumers selected more expensive mortgages that had a lower broker fee because they focused on the fee rather than the total cost.

Challenges Policymakers Face in Correcting Consumer Bias

  • Insufficient knowledge. The information needed to diagnose consumers’ bias can be difficult to acquire. Identifying consumer preferences, which may conflict, can be challenging for policymakers. Additionally, measuring bias can be imprecise or even impossible.
  • Behavioral bias. Policymakers’ judgment may be subject to the same cognitive biases that consumers face. For example, policymakers may be subject to action bias, which pushes people to take action when faced with risk and uncertainty. Additionally, they may be subject to confirmation bias—the tendency to interpret information in a way that confirms one’s prior assumptions about the world.
  • Incentives. Policymakers cannot care as much about consumers’ well-being as the consumers themselves. Moreover, regulatory agencies may overestimate consumers’ bias in order to maximize demand for their services and regulations.


An action is considered paternalistic if and only if

  • it aims to influence the consumer’s behavior,
  • it aims to improve the consumer’s welfare,
  • it substitutes the paternalist’s judgment for the consumer’s, and
  • it is undertaken on the ground that the paternalist’s judgment is superior to the consumer’s.

Additionally, a paternalistic action can be either hard (coercive) or soft (noncoercive).

While most disagreements over nudges tend to focus on the first two conditions, this paper looks at the last two conditions. Nudges need not question or replace the consumer’s judgment, but can instead counter behavioral biases by aiding consumer decision-making.

Paternalistic Tools

Paternalistic tools that policymakers can use include changing the default option, using reference points, and exploiting loss aversion.

  • Default options. For example, consumers once had to opt out of overdraft protection, which allows bank customers to get cash at an ATM (with a charge) even if there is no money in their account. While there may be times when consumers want this convenience, many consumers reported overdrawing their accounts by mistake. In an attempt to remedy this, the Federal Reserve Board changed the default so that consumers had to opt in to this service.
  • Reference points. In New York City, taxicabs began recommending tips of between 20 and 30 percent, anchoring consumers’ expectation of a reasonable tip to 20 percent and thus increasing the average tip amount from 10 percent to 22 percent.
  • Loss aversion. In one experiment, teachers were given a bonus up front rather than at the end of the year, and the money was taken away at the end of the year if their performance did not meet expectations. Students of these teachers achieved higher math scores than students whose teachers received a bonus at the end of the year.

Nonpaternalistic Tools

Nonpaternalistic interventions, such as providing timely feedback, increasing the popularity of choosing certain options, and simplifying processes, can be as effective as paternalistic tools at helping consumers make better choices. And if policymakers have misdiagnosed the behavioral bias influencing consumers, nonpaternalistic policies do not impose policymakers’ errors on consumers.

  • Reminders. Nonpaternalistic interventions such as reminders become particularly important in cases such as medication nonadherence, where patients with chronic illnesses forget to take their medications, because these problems have a variety of behavioral and nonbehavioral causes.
  • Priming. For some choices, policymakers can attempt to increase the salience or popularity of a certain option. For example, priming can raise the salience of specific future outcomes: if consumers are told to think of their older selves when making decisions about their financial future, they may be more likely to save.


Efforts to nudge Americans to make better choices about their finances or health often run into opposition owing to their paternalistic nature, but nonpaternalistic nudges can be equally effective at improving consumers’ choices. In contrast to paternalistic nudges, nonpaternalistic policies do not impose policymakers’ errors on consumers if policymakers misdiagnose the underlying behavioral bias, and they thus avoid harming consumers by pushing them toward suboptimal choices.