ICES Seminar in Experimental Economics and Game Theory
Managing Crowding Out: How Lawmakers Can Avoid Unintended Incentive Effects
Friday, August 30, 2024 12:00 PM to 1:00 PM EDT
Vernon Smith Hall (formerly Metropolitan Building), 5183
The ICES Seminar in Experimental Economics and Game Theory of the Fall 2024 semester will feature:
Antonin Scalia Law School, George Mason University
Managing Crowding Out: How Lawmakers Can Avoid Unintended Incentive Effects
Abstract
Monetary rewards and fines are important tools in lawmakers’ arsenal to encourage desirable and discourage undesirable behaviors. However, rewards and fines can crowd out moral motivation, dampening and sometimes even reversing the intended effect. For example, encouraging people to donate blood by paying them has been found to lead to less, not more, blood donors.
Laboratory and field studies in the last decades have aimed to identify and explain crowding-out effects in different social contexts. These studies have focused on scenarios of purely pro-social motivation. Take studies on blood donation as an example. Donating blood benefits recipients but does not benefit donors. However, in many situations of great policy relevance, people act out of both self-interested and pro-social motivations. An example of the Covid-19 pandemic is vaccination against a contagious disease. One may choose to receive vaccination to protect oneself AND to prevent spreading the disease to others. If policymakers provide monetary incentives to vaccinate, as some countries did during the pandemic, should we expect positive incentive effects to be offset by crowding out? We will show in the study that if we use the well-studied crowding out effects of purely pro-social motivated behavior as a policy benchmark, we will systematically overestimate the impact of crowding out effects for mixed-motive behavior.
To answer this question and help policymakers predict in what type of situation monetary interventions may backfire, we need a better understanding of the mechanisms that drive the crowding-out effect. Our paper focuses on two often-discussed theories. Social signaling theory (SST) suggests that individuals derive utility from being perceived as prosocial actors by others. Once monetary incentives are introduced, it becomes difficult for them to signal to others that their actions are based on prosocial motivations; the monetary incentives thus crowd out the utility derived from being perceived as a prosocial actor (Arieli et al. 2009). Self-determination theory (SDT), on the other hand, claims that monetary incentives crowd out motivation because individuals value autonomy, and when monetary incentives are used to interfere with their autonomy, they weaken intrinsic motivation (Deci et al 2010).
We conduct two experiments to analyze motivational crowding out of mixed-motive behavior and test which of the two theories may better explain the effect. In our first experiment, participants in the Baseline treatment are presented with the scenario of a contagious disease and are asked whether they want to vaccinate against that disease. In a second scenario, they are asked whether they would be willing to donate blood, a purely prosocial behavior. For the same two scenarios, the treatment study introduces monetary incentives: participants in the treatment group are asked the same question after being told that monetary incentives are provided to encourage vaccination and blood donation.
Both theories would predict crowding out in the blood donation scenario - either because monetary incentives would prevent participants from signaling prosocial motivations or because they would interfere with participants’ autonomy - and we in fact do observe crowding out in Incentive treatment: When incentives are provided, our subjects indicate a significantly lower willingness to donate compared to the Baseline.
For the vaccination scenario, the predictions of the theories are different: signaling theory predicts that we should not see crowding out. Subjects are already prevented in the baseline from signaling a social motivation for vaccination because an observer cannot tell whether someone is vaccinated out of a self-interested motivation to protect themselves or out of a pro-social motivation to protect others. Thus, providing incentives that suggest an alternative self-interested motive for vaccination can only confirm that subjects may not be pro-socially motivated. Self-determination theory, on the other hand, predicts crowding out: the introduction of monetary incentives interfere with the person's perceived decision autonomy and would therefore crowd out prosocial motivation relative to the baseline.
Our data show almost no crowding out in the vaccination treatment, but strong crowding out in the blood donation scenario. Thus, our finding in the first experiment provides support for the social signaling theory and suggests that incentives may not backfire in mixed-motivation activities such as vaccination.
However, the data do not rule out the possibility that vaccination may be a case where crowding out never occurs, for example because self-interested reasons may dominate most people's vaccination choices. We therefore need to measure people's prosocial motivation for vaccination and then show when it is not crowded out, as in our first study, and when it can be crowded out.
In our second experiment, we directly test our theory that the difficulty of signaling prosociality prevents incentives from crowding out mixed-motive behavior by experimentally turning on and off subjects' ability to signal their prosocial motivation. In a 2x2 design, we vary whether participants have a personal health reason for getting vaccinated that is either public, so everyone knows or private, so only they know themselves and whether participants are offered incentives to vaccinate or not.
By making the disease harmful only to women (if the participant is male) or only to men (if the participant is female) our design ensures that subjects can signal a pro-social motivation. For example, each female participant is told that while both genders can spread the disease only men suffer from severe symptoms; thus women’s motivation to vaccinate is purely prosocial. And because gender is usually observable to others, women can signal their motivation. Similarly, each male participant is told that he could vaccinate to protect women, but he does not suffer any harmful effects from the disease.
Second, we instruct subjects that they also have a self-interested motivation to vaccinate: they were previously infected with a dangerous virus, which puts them at risk if they are infected again. In the public condition, this prior infection is known to the health department and subjects are vaccinated in the group of people who face own health risks if infected.
In the private condition, they know of their prior infection but the health department does not and so they are vaccinated in the group of people who have no health risks. Thus, a subject who privately knows that she was previously infected may signal a prosocial motivation when she gets vaccinated. By contrast if her personal risk is public and she is given priority for vaccination, she cannot. In the second treatment arm, we provide incentives for vaccination in both the public and the private conditions.
Social signaling theory would predict that in the public conditions, incentives should not cause a crowding-out effect, since participants cannot signal a prosocial motivation. By contrast, there should be crowding out effects in the private treatment arm, where subjects can signal prosocial motivation without incentives but not when incentives are provided in the treatment. Self-determination theory, on the other hand, predicts crowding out effects in both the private and the public treatment arms, since only the incentive condition interferes with perceived autonomy. Our experimental findings here again support social signaling theory.
Our study provides policy makers with insights to improve their management of crowding out effects by showing that in a broad category of situations in which peoples` actions can be driven by both self-interested as well as prosocial motivations, monetary interventions are unlikely to cause crowding out effects. And second, our evidence contributes to the understanding of crowding out mechanisms by suggesting that the social signaling theory provides a better explanation for crowding out effects of mixed motive behavior than self-determination theory.
For more information about the Seminar Series, please visit the Seminar Schedule homepage.
