A study on the motives behind paternalistic behavior to what degree are people willing to act paternalistically, by interfering with another person’s autonomy, with the intent to increase this person’s well-being?

(1)

A study on the motives behind

paternalistic behavior

To what degree are people willing to act paternalistically, by interfering with another person’s autonomy, with the intent to increase this person’s well-being?

Nol Gulcher (11965681)

Master thesis Managerial Economics & Strategy

15 ECTS

Supervisor: dr. Sneha Gaddam

Abstract

This paper examines the motives behind paternalistic behavior. In particular, I examine how the willingness to act paternalistically depends on the impact one can make on another person’s well-being. In previous literature, it is assumed that when individuals make a paternalistic decision, there exists a moral trade-off between other’s well-being and the other’s autonomy. When making paternalistic interferences people face a trade-off between these two moral values. Whether a person interferes will reveal his moral preferences. In this paper, it is hypothesized that when a person has an informational advantage, the willingness to act paternalistically increases. Furthermore, I expect that people are more willing to act paternalistically when the impact, that a person can have on another’s well-being, increases. The main findings of this experiment are that there is a significant increase in the willingness to act paternalistically when a person has an informational advantage and that the willingness to act paternalistically is significantly larger for men, compared to women. Finally, I found no apparent effect of the impact that a person can have on another person’s well-being on the willingness to act paternalistically.

(2)

Statement of Originality

This document is written by Nol Gulcher who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Acknowledgement

I would like to thank my thesis supervisor dr. Sneha Gaddam for all of her valuable feedback and support. Whenever I had a question I could always approach her and rely on her advice. Thank you. I would also like to thank Joost Verdonkschot, Janneke Janssen, and Mirjam v.d. Gulik for providing me with the opportunity to conduct my experiment at ‘Het Spinoza Lyceum’ in Amsterdam. Without their support, I wouldn’t have been able to perform my analysis. Finally, I would like to thank my parents and my girlfriend for their ongoing support throughout my studies and through the process of writing my thesis.

(4)

1. Introduction

Paternalism can be defined as the interference of the government or individuals with another person’s autonomy to make choices of their own, without his or her consent, and motivated by claiming that it is in his or her best interest (Dworking, 2002). The role of paternalistic behaviour has great significance in public policymaking, law, professional-client relationship, and interpersonal relationship. Under what circumstances is it permissible for the state or individual to intervene in the lives of another person for their own good?

In recent years the literature on paternalistic interventions received great interest from public policymakers and academics. Especially after Thaler and Sunstein (2008) published their book ‘Nudge- Improving decisions about health, wealth and happiness’ policymakers, mainly from the US and UK, were influenced in their approach on areas including pensions savings, public health, and education (Thaler 2015). Inspired by the work of Richard Thaler and Cass Sunstein, the recent literature on paternalism takes a narrow focus on ‘soft’ paternalism, mainly referred to as Libertarian Paternalism, which is an approach on paternalism that permits public and private institutions to nudge people to make the right decision for themselves, while preserving their freedom of choice (Thaler and Sunstein, 2003). Two fields where nudges have had a crucial impact are retirement savings (Choi et al. 2004; Thaler & Bernatzi, 2004) and organ donation systems (Johnsen and Goldstein, 2003).

However, Libertarian Paternalism only addresses a limited part of the concept of paternalism. In addition, the existing literature is mainly focused on the appropriate role of paternalistic behaviour in public policy making and takes a narrow approach to paternalism. It is important to make a distinction between broad and narrow paternalism. A narrow paternalist is solely interested in the appropriate role of state intervention, for instance the use of legal coercion, whereas a broad paternalist is interested in paternalistic behaviour by the state, institutions (e.g. hospital policy), and by individuals (Dworking, 2002). On the nature of paternalistic behaviour on an individual level, there is only limited research available.

A recent study by Altenau and Jenssen (2017), contributed to the understanding of paternalistic behaviour on an individual level. The authors test the moral motivation behind hard paternalistic interventions by conducting an incentivized experiment. In this experiment, a moral trade-off was created where an observer had to decide whether a stakeholder should be allowed to make a decision that only affected the stakeholder’s payoff. The authors examined

(6)

the causal impact of information asymmetry on the willingness to act paternalistic and found an increase in paternalistic interventions when the observer has an informational advantage. However, the majority of observers choose not to intervene. The results indicate that people are morally motivated by the well-being of others but also have a strong aversion against interfering with another person’s autonomy. In addition, the authors found that men are more likely to act paternalistically compared to women. It raises an interesting question under what circumstances people are willing to interfere with another person’s autonomy, with the intent to promote his or her well-being.

Altenau and Jenssen (2017) assume that when individuals make a paternalistic decision, there exists a moral trade-off between other’s well-being and other’s autonomy. However, the authors did not examine the sensitivity between these moral values. It would therefore be interesting to examine how the willingness to act paternalistically depends on the impact one can make on another person’s well-being. The purpose of this study is to contribute to the understanding of the nature of paternalistic behaviour in interpersonal relationships by examining the sensitivity of the trade-off between these moral values, and by replicating the findings by Altenau and Jensen (2017).

It is well established in previous research that people value the well-being of others (Charness & Rabin, 2002; Fehr & Fischbacher, 2002), and that individuals value autonomy beyond their instrumental benefit (Barling et al., 2014; Fehr et al., 2014). In economics, how individuals value other’s autonomy and other’s well-being has mainly been studied in the context of the incomplete contract literature by analysing concepts such as decision-making, authority, and ownership in institutions and organizations (Simon, 1951; Grossman & Hart, 1986; Aghion & Tirole, 1997). These studies focus on how individuals make decisions in a hierarchical relationship. For instance, the relation between employer and employee, or between the state and its citizens.

The established literature mainly focuses on the nature behind paternalistic behaviour for private and public policy designing. This study will move beyond the scope of the existing literature, taking a broader approach by examining hard paternalistic interferences, in a non-hierarchical relationship. This study will focus on the reason behind paternalistic behaviour when people are forced to make a decision between two moral values. When people have to make decisions, it is well established that they have moral preferences and are morally

(7)

motivated (Kahneman et al., 1986; Fehr & Camerer, 2002). However, what drives these moral decisions is still a complex field.

When making hard paternalistic interferences people face a trade-off between these two moral values. Namely, another person’s autonomy and another person’s well-being. Whether a person interferes will reveal his moral preferences. This study will provide empirical evidence on the sensitivity of the trade-off between these moral values, and aims to answer the

following questions:

To what degree are people willing to act paternalistically, by interfering with another person’s autonomy with the intent to increase this person’s well-being?

and

How does the willingness to act paternalistically depend on the impact one can make on another person’s well-being?

A situation is created where people face a moral trade-off between another person’s autonomy and his well-being. In this experiment, an observer will have to decide whether a stakeholder should be allowed to make a decision that only affects the stakeholder’s own payoff. The observer can let the stakeholder decide for himself or he can overrule his decision right and decide on the stakeholder’s behalf. By introducing information asymmetry between the observer and stakeholder, the observer will be better informed on the stakeholder’s payoffs. In this way, I can adjust the impact that the spectator can have on the stakeholder’s well-being. First, this study analyses the impact that the information asymmetry has on the willingness to act paternalistically. Second, I will change the level of impact that the spectator can make on the well-being of the stakeholder by changing the pay-off scheme of the stakeholder. The latter will allow me to analyse the sensitivity of the trade-off between how people value autonomy and another person’s well-being.

The main findings of this experiment are that there is a significant increase in the willingness to act paternalistically when the observer has an informational advantage, and that the

willingness to act paternalistically is significantly larger for men, compared to women. Also, I analysed whether the impact that a person can have on another person’s well-being, by acting paternalistically, positively affect the willingness to act paternalistically. I found no apparent

(8)

effect. Finally, the results indicate that in general, people value another person’s well-being above his or her autonomy, as the majority of subjects decided to act paternalistically when they were in the position to improve the stakeholder’s well-being.

This paper is structured as follows. In section 2, I present the existing literature on paternalism, and two moral values, autonomy and well-being. In section 3, I describe the methodology, experimental design, and predictions. In section 4, the main results are presented. Finally, in section 5, the main findings and limitations are discussed.

(9)

2. Literature Review

2.1 Defining paternalism

One of the most commonly cited definitions of paternalism has been provided by Gerald Dworkin (2002). He defines paternalism as: “the interference of a state or an individual with another person, against their will, and defended or motivated by a claim that the person interfered with will be better off or protected from harm”. As the concept of paternalism is normally attached to actions, there are three conditions for an act to be paternalistic (Dworkin, 2002; Grill, 2011). Following the definitions by Dworkin (2002) and Grill (2011), person X acts paternalistically towards person Y when his action meets the following conditions:

1. Interference Condition: person X interferes with the liberty or autonomy of person Y 2. Consent Condition: person X does so without the consent of person Y

3. Benevolence Condition: person X does so because he is motivated, and perhaps also justified, by the good of person Y

In the following section, the three conditions are further explained.

2.1.1 Interference condition

The first condition implies that paternalistic behaviour involves limiting a person’s liberty or autonomy (Dworkin, 2002; Grill, 2011). It is important to make a clear distinction between interference with liberty and autonomy. Following the terminology of Berlin (1969), liberty refers to the freedom to make choices without interference from a third party, and autonomy refers to free will.

Liberty, in the context of paternalism, can be best understood as the availability of a range of different options to a person (Grill, 2011). Thus, by limiting another person’s liberty, the range of different options becomes smaller. For example, the state implementing a mandatory social security and medical coverage system. This will limit the available options as it is not possible to have no social security.

(10)

Autonomy, in the context of paternalism, can be seen as self-determination. According to Grill (2011), the interference with another person’s autonomy can be seen as the substitution of a person’s judgment or agency. In this case, the paternalist acts or judges in place of the subject.

The following example illustrates the difference between liberty and autonomy. The state imposes a legal alcohol limit to drive a car. I am not at liberty to drink more than the imposed alcohol limit. However, I have the option to drink more than I am allowed if I am willing to risk getting punished. In this case, the state interferes with my liberty but not with my autonomy to choose how much I drink before driving. In contrast, imagine I have an alcohol interlock in my car that restricts me from driving under influence, imposed by the state. In this case, the state interferes with both my liberty and my autonomy as I am now unable to drink and drive.

In this study, an observer has to decide whether a stakeholder should be allowed to make a decision that only affects the stakeholder’s payoff. By interfering, the observer will interfere with the stakeholder’s autonomy, as the observer will overrule the decision made by the stakeholder.

2.1.2 Consent condition

For an interference to be paternalistic, it must be that it is done without the consent of the stakeholder (Dworkin 2002, Grill 2011). This consent is only valid if it is broadly voluntary and the product of reasoned choice (VandeVeer, 1986). For example, if an alcoholic asks his friend to hide his car keys, to prevent him from driving under influence, and his friend does so accordingly, he does not act paternalistically by doing so. In this case, the friend was asked to hide the keys with the consent of the alcoholic. Therefore, the stakeholder’s autonomy has not been compromised because he voluntarily consented to the interference by his friend.

2.1.3 Benevolence condition

This condition is aimed at the agent who is interfering with the liberty of autonomy of the stakeholder. The foremost motivation for interfering with the stakeholder must be to increase his well-being or to prevent him from harm (Grill, 2011). This does not mean that this is actually the case. However, as long as the agent is benevolently motivated to interfere with stakeholder, the interference can be seen as paternalistic. The latter is mainly aimed to

(11)

promote the stakeholder’s physical health, or to promote his financial situation (Grill, 2011). In this thesis, the observer acts benevolently by maximizing the stakeholder’s expected payoff.

2.1.4 Hard vs soft paternalism

In this study, I focus on hard paternalistic interferences by individuals in a non-hierarchical relation. The literature on paternalism makes a major distinction between hard and soft paternalism (Feinberg, 1989; Pope 2004). This distinction refers to the degree to which the individual concerned with is acting voluntarily or autonomously (Le Grand & New, 2015). When an individual is acting voluntarily (or autonomously), a hard paternalist knows that by interfering with another person’s decision making he will compromise his autonomy. Yet, by limiting his autonomy to promote his well-being, a hard paternalist believes there is an acceptable trade-off between these moral values (Le Grand & New, 2015).

An individual is acting voluntarily (or autonomously) when he acts like a rational agent, meaning that the individual concerned with is making judgments, or decision, in his or her best interest. However, similar to many of the concerns in behavioural economics, soft paternalists believe that in many situations individuals lack the reasoning to make the right decision. Studies on behavioural economics suggest that people suffer from bounded rationality (Simon, 1972), and are influenced by heuristics (Tversky and Kahneman, 1974) that lead to systematic biases in our ability to make decisions. In addition, we suffer from self-control problems and reference dependence due to nonstandard preferences and nonstandard beliefs (DellaVigna, 2009). When these anomalies are present in our decision making, soft paternalism assumes people are acting in a nonvoluntary manner. Therefore, interfering with another person’s decision making doesn’t offend his autonomy since his decision is not really his own (Le Grand & New, 2015).

In recent years the literature on paternalistic interventions received great interest from public policymakers and academics. A more practical approach to the trade-off between individuals' well-being and their autonomy was introduced as "Libertarian" paternalism by Richard Taler and Cass Sunstein (Taler & Sunstein, 2003; Taler & Sunstein, 2008) and "asymmetric" paternalism by Colin Camerer (Camerer et al., 2003). The underlying idea to both concepts is simple. Influenced by studies in behavioral economics, on how human decision making is

(12)

systematically biased, as mentioned before, the authors propose a subtler approach to

interfering with people their decision making. Through the concept of nudges, policy makers can actively manipulate reasoning failure to ‘nudge' people to make the right decisions for themselves by changing the conditions under which people make decisions rather than interfering with the decision making itself. The focus is on the design of the environment where choices take place. By doing so the impact on autonomy is minimized as the measure of individual choice is preserved (LeGrand & New, 2015). Two fields where these concepts of soft paternalism have had a crucial impact are retirement savings (Choi et al. 2004; Thaler & Bernatzi, 2004) and organ donation systems (Johnsen and Goldstein, 2003).

Although the impact of the proposed concepts of soft paternalism has been significant on private and public policy making, it only addresses a limited part of paternalism as proposed by Dworking (2002). In addition, one could argue that the interference condition has not been met. In hard paternalistic interferences, the paternalist acknowledges that the autonomy and freedom of choice of the person interfered with will be compromised but believes that the trade-off between autonomy and the person’s well-being is acceptable (LeGrand & New, 2015).

2.2 Moral trade-off

This study analyses the sensitivity between two moral values when individuals are in the position to make hard paternalistic interferences in a non-hierarchical position. It is well established in the literature of behavioural economics that people are not solely self-interested and are morally motivated when making decisions (Kahneman et al., 1986; Fehr &

Fischbacher, 2002). People care about others well-being and are motivated by reciprocity (Charness & Rabin, 2002). In addition, people value autonomy beyond its instrumental value as they assign a positive intrinsic value to hold autonomy (Deci & Ryan, 1985; Bartling et al., 2014).

A hard paternalist knows that by interfering with another person’s decision making he will compromise his autonomy. However, he believes that there is an acceptable trade-off between limiting his autonomy to promote his well-being (Le Grand & New, 2015). Based on the evidence that people are morally motivated when making decisions, this study will investigate the sensitivity between respecting other’s autonomy and other’s well-being. Altenau and

(13)

Jensen (2017) tested this moral trade-off by creating a situation where a spectator had to decide whether a stakeholder should be allowed to make a decision that only affected the stakeholder’s payoff. The authors found an increase in paternalistic interventions when the spectator has an informational advantage. However, the majority of spectators choose not to intervene. In addition, they found that men are more likely to act paternalistically compared to women. The results indicate that people are morally motivated by the well-being of others, but also have a strong aversion to interfering with another person’s autonomy.

2.2.1 Autonomy

Personal autonomy can be best understood as the ability to make own decisions, to live one’s life according to reason, and without being influenced by manipulative and distorting external forces (Christman, 2008). Within the context of paternalism, autonomy is best understood as the right to decide. Hence, when making a paternalistic interference this involves a lack of respect for autonomy (Christman, 2008).

According to social psychologists, people intrinsically value autonomy. Deci and Ryan (1985), who proposed the self-determination theory, hypothesize that autonomy is “essential for ongoing psychological growth, integrity, and well-being” (Deci & Ryan, 2000). The need for autonomy reflects the desire by people to be in control of one’s own decisions. Being in control of one’s own decisions raises happiness and motivation as it increases the perceived procedural utility when making choices (Frey et al. 2004).

Hence, hard paternalistic interventions do not only impact an individual’s autonomy, but also affect his or her intrinsic motivation (LeGrand & New, 2015). Accordingly, people are likely reluctant to act paternalistically when they place a high value on autonomy. This is consistent with the findings from Altenau and Jensen (2017). Although the spectators were in a better position to promote the well-being of the stakeholders, the majority choose not to intervene and thereby allowing the stakeholders to make choices for their own. Another study, by Bartling et al (2014), tested how individuals value decision right compared to its instrumental value. They found that individuals value the right to make own decisions beyond its

instrumental value. According to the authors, the size of the intrinsic value of decision right depends on the stakes. Doubling the stakes almost doubled the intrinsic value.

(14)

These findings are also consistent with the outcome of an experiment conducted by Fehr et al. (2013). The authors studied the motivation and incentive effect in an authority delegation game. In this game, the stakeholder had to decide which project to implement. The

stakeholder could make this decision for himself or delegate the decision right to an observer. The experiment consisted of different treatments varying in the level of information

asymmetry with respect to the outcomes of the different projects. The main finding is that individuals often retain authority even if it is in their best interest to delegate decision rights.

Knowing that individuals are morally motivated and value autonomy intrinsically, Altenau and Jensen (2017) suggest that people do not only care about their own autonomy, but also have preferences to respect other people their decision rights. Hence, when faced with a moral trade-off between another person’s autonomy and his well-being, one could expect that

people are reluctant to act paternalistically even if they could improve this person’s well-being.

2.2.2 Well-being

A crucial feature of the concept of paternalism should be that it is intended to improve the good of the person whose autonomy is interfered with (LeGrand & New, 2015). In this case, the ‘good' refers to the well-being of the person in question by preventing harm or to promote benefit (Feinberg, 1990). It is well established in previous research that people value the well-being of others.

The standard economic theory builds on the assumption that all people are solely interested in their own material payoff. However, a breadth of empirical evidence suggests that this

assumption is an overestimation of the behavior observed in reality (Charness & Rabin, 2002; Fehr & Fischbacher, 2002). Common explanations used in economic literature to explain non-selfish behavior in decision making are people’s social preferences and other-regarding preferences, like altruism (Simon, 1993). Accordingly, Charness and Rabin (2002) found that people are willing to take actions, which are costly to their own payoff, with the sole purpose of increases the payoff of others.

(15)

The evidence on social preferences suggests that people care about others’ well-being. This suggests that, when faced by a moral trade-off between another person’s autonomy and his well-being, on might be willing to interfere with his autonomy, if they are able to increase the well-being of the person in question.

Altenau and Jensen (2017) tested this moral trade-off and their results indicate that people are morally motivated by the well-being of others, but also have a strong aversion against

interfering with another person’s autonomy. When making hard paternalistic decisions, people face a moral trade-off between these two moral values, another person’s autonomy and his well-being. This study will explore the sensitivity of this trade-off.

(16)

3. Methodology

This study will analyse the motives behind paternalistic behaviour by looking at the degree to which people are willing to interfere with another person’s autonomy, with the intent to promote his or her well-being. Research has shown that people value both the well-being and autonomy of others. Hence, they are confronted by a moral trade-off when making hard paternalistic decisions.

To analyse how individuals behave when confronted with this moral trade-off, I will conduct an incentivized experiment with high-school students. The study has a Within-Subject Design that will allow me to investigate the sensitivity of how individuals value other’s well-being and other’s autonomy. A "Within-Subject Design" has been applied to maximize the number of responses per subject. Also, since each individual is exposed to all treatments, by

examining how individual behaviour differs when the experimental circumstances change, causal estimates can be obtained (Charness et al., 2012). The experiment takes place in a high-school with pre-selected classes by the school. This makes it difficult to completely randomize the assignment of the experiment. However, by using a within-subject design this will not be a problem as the internal validity does not depend on random assignment

(Charness et al., 2012).

Altenau and Jensen (2017) were the first to experimentally test which moral preferences are dominant when faced with a trade-off between another person’s autonomy and his well-being. Similar to the study of Altenau and Jensen (2017), this study analyses the impact of information asymmetry on the willingness to act paternalistic and how individuals value the well-being and the autonomy of others. However, the authors did not examine the sensitivity of the trade-off between these moral values. This paper builds on the existing literature by replicating the findings by Altenau and Jensen (2017), and by examining the sensitivity of the trade-off between these moral values.

I will conduct an incentivized experiment. An environment is created where the participants face a moral trade-off. In this experiment, an observer will have to decide whether a stakeholder should be allowed to make a decision that only affects the stakeholder’s own payoff. The observer can let the stakeholder decide for himself or he can overrule his decision right and decide on the stakeholder’s behalf. A situation is created where the observer is better informed on the stakeholder’s payoff. Hence, there is information asymmetry present between the

(17)

stakeholder and the observer. This allows me to analyse the impact of information asymmetry on paternalistic behaviour. In addition, throughout the different treatments, I will vary how much impact the observer can have on the well-being of the stakeholder. This will allow me to examine the sensitivity of the trade-off between how people value both the well-being and the autonomy of others.

3.1 Experimental Design

To address the research questions, I will conduct an incentivized experiment with high-school students. Similar to the research by Alteneau and Jensen (2017), the experiment consists of three stages and two roles, the stakeholder, and the observer. The first stage of the experiment was conducted a week prior to the second stage. Therefore, the stakeholders and observers did not interact with one another. In the following sections, the experimental stages, the different roles, and the experimental treatments will be further explained.

3.1.1 Experimental Stages

In the first stage, the stakeholder performed a simple task, matching 10 national flags to their respective country. Each stakeholder received a participation fee of €5, - and an additional bonus payment. After completing the task, the stakeholder had to decide between two

different payment options for his bonus. His decisions will stay anonymous. In addition, they are informed that a second person, the observer, will be given the opportunity to overrule the decision made by the stakeholder and decide which bonus option the stakeholder will receive. This stage is solely used as an instrument for the second stage to ensure that the decisions made in the second stage have real-life consequences.

In the second stage, the observers are paid a fixed participation fee, will go through all different treatments, and in each treatment, they have two options. The first option will allow the observer to overrule the decision made by the stakeholder in the previous round, and he will decide which bonus option the stakeholder will receive. Alternatively, if the observer chooses the second option he will allow the stakeholder to make his own decision, and thereby respect the stakeholder’s autonomy. The preferences of the stakeholder will not be informed to the observer.

In the third stage, the choices made by the stakeholders and the observers will be randomly matched, and the stakeholders will be paid according to the observers' decisions. Due to

(18)

financial constraints, a lottery will choose a total of five participants who will be paid. Their payoff depends on their own decisions, the decisions made by the observers, and the outcome of the lottery.

3.1.2 Experimental Roles

In this experiment, depending on the experimental stage, the participants were assigned to two different roles, the stakeholders or the observers. The stakeholders are included in this

experiment to ensure that the decisions made by the observers have real-life consequences. The decisions made by the stakeholders serve as an instrument for the second stage, and their actions are therefore irrelevant for the analysis of this experiment.

The Stakeholders

A questionnaire was conducted, and in total 10 stakeholders were recruited who each had to complete a simple task, matching 10 national flags to their respective country. The

stakeholders were paid a participation fee of 5 EUR and an additional bonus payment. The stakeholders had to decide between two different bonus options. The stakeholders were informed that the bonus payment depended on their own choice, the observer’s choice, and the outcome of a lottery. The choices made by the stakeholders will be anonymous. The amount of the bonus payment could range between 0 EUR to 20 EUR, depending on the treatment they are assigned to, the decisions of both the stakeholder and the observer, and the outcome of a lottery.

The choices made by the stakeholders are irrelevant for the analysis of this study and serve as an instrument for the second stage.

The observers

I conducted a controlled experiment at ‘Het Spinoza Lyceum’, a high-school in Amsterdam. In total 101 vwo-students (pre-university education) participated in my experiment. All students were between 16 and 18 years old. Due to the high level of similarity among the observers, I will only compare the results of the treatment effects between gender.

The students were informed that by participating they could earn a participation fee, would have to answer multiple questions concerning restricting another person’s freedom, and finally had to answer a background question about their gender.

(19)

When participating, I emphasized that the decisions of the observers could have real-life consequences for another person’s monetary payoff. They are informed that a stakeholder had to match 10 national flags to their respective country and that the size of the stakeholder’s payoff consisted of a participation fee and an additional bonus payment. Also, the observers are informed that their decisions will stay anonymous.

First, the observers were fully informed on the payment options that the stakeholders received and that the stakeholders are aware that a second person, the observers, are given the

opportunity to overrule their decision and decide on their behalf. Hence, the observers are fully aware of what information the stakeholder does and does not have. Therefore, if any information asymmetry is present, the observer should be aware of this.

Second, the observer had to decide whether to allow the stakeholder to decide between the two different bonus options by himself. He can either respect the stakeholder’s autonomy by letting the stakeholder decide for himself, or the observer can overrule the stakeholder’s decision and decide on the stakeholder’s behalf which bonus option he will receive. Finally, the observer is asked to answer a background question about his or her gender.

The decisions made by the observer has no consequences for his payoff as he can only impact the pay-off (well-being) of the stakeholder. Therefore, the observer faces a pure moral

dilemma when deciding whether to interfere or not.

3.1.3 Experimental Treatments

In this experiment, there are five different treatments which vary in the level of information asymmetry and in the level of impact that the observer can have on the stakeholder’s well-being.

First, I tested the willingness to act paternalistically when there was no information asymmetry present. The observer and the stakeholder were equally informed on the bonus options, and this treatment served as the base treatment. Second, information asymmetry in favour of the observer was introduced. In this treatment, the observer was better informed on the specifications of the bonus options. This created a state where the observer was in a better position to make a decision on which bonus option would maximize the expected utility for the stakeholder. The observer faces a moral dilemma as he is able to improve the well-being of the stakeholder, by interfering with his autonomy. This allows me to make causal

(20)

interpretations on the impact of information asymmetry on the willingness to act

paternalistically. Finally, an informational advantage in favour of the observer is present in the last three treatments, similar to the second treatment. However, the level of impact that the observer can make on the well-being of the stakeholder is manipulated across treatment three to five.

In the following sections, the base treatment and the treatment manipulations are elaborated.

The base treatment

In the first treatment, the base treatment, there is no information asymmetry present between the observer and the stakeholder. Both are equally informed on the bonus options.

In stage one, the stakeholder performed a simple task and his payoff consisted of a

participation fee and an additional bonus payment. After completing the task, the stakeholder had to choose between two bonus options, alternative A or alternative B. Alternative A leads to a bonus payment of 5 EUR with a probability of 100%, and alternative B leads to a bonus payment of 20 EUR with a probability of 50%. Thus, alternative B leads to the highest expected utility and is clearly the better option. Finally, the stakeholder is informed that a second person, the observer, will decide whether he can choose the payment option by himself/herself, or whether the observer will make the decision on his behalf.

In the second stage, the observer is informed that the stakeholder has the same information with respect to the bonus options. Therefore, the observer is informed that there is no information asymmetry present. In this treatment, there is no moral trade-off present. The stakeholder will choose the alternative that is in his/her best interest and the observer does not have to interfere with his/her autonomy to increase his/her well-being. Therefore, I assume that the observer will not act paternalistically in this treatment.

The treatment manipulations

In the following treatments, information asymmetry in favour of the observer is introduced and the impact that the observer can have on the well-being of the stakeholder will be

manipulated. Therefore, the base treatment serves as a control treatment which will allow me to examine the impact of these manipulations on the willingness to act paternalistically.

(21)

Observer informational advantage

In the second treatment, the observer will have an informational advantage. The observer will have more information, compared to the stakeholder, on the outcomes of the bonus payment alternatives. In this treatment, the alternatives for the bonus payment options are similar to the base treatment. However, the stakeholder does not know the probabilities.

Similar to the first treatment, in the first stage, the stakeholder had to choose between two alternatives for an additional bonus payment. However, in the second treatment, the

stakeholder has no information about the probabilities of receiving alternative B. Hence, the stakeholder can decide between receiving alternative A, 5 EUR with a probability of 100%, or alternative B, 20 EUR with an unknown probability.

In the second stage, the observer is informed that the stakeholder did not receive any information about the probability of receiving alternative B. The preferences of the

stakeholder will not be informed to the observer. By revealing the probability of receiving alternative B to the observer only, I created a situation where the observer has an

informational advantage. Similar to the base treatment, alternative B will lead to the highest expected utility. However, the observer is unaware of the level of risk aversion of the stakeholder as he or she might prefer a certain 5 EUR over alternative B with an unknown probability, or vice versa.

When making paternalistic interferences one can never be completely certain about the preference of the person interfered with because he has not given his consent to do so and revealed his preferred bonus option. Hence, uncertainty about the stakeholder’s risk aversion has been included in this experiment.

The observer has more information about the possible outcomes of the bonus payment and is in the position to improve the stakeholder’s well-being, by interfering with his autonomy. Whether a person interferes will reveal his moral preferences.

Impact on well-being

The impact that the observer can make on the well-being of the stakeholder is the difference of the expected payoff between the two alternatives for the bonus payment.

(22)

In the third treatment, the experimental circumstances are similar to the previous treatment, apart from the probability of receiving 20 EUR for Alternative B. In this treatment, the stakeholder had to choose between two bonus options, alternative A and alternative B. Alternative A leads to a bonus payment of 5 EUR with a probability of 100%, and alternative B leads to a bonus payment of 20 EUR with a probability of 30%. Thus, alternative B leads to the highest expected utility.

The observer has more information about the possible outcomes of the bonus payment and is in the position to improve the stakeholder’s well-being, by interfering with his autonomy. However, compared to the previous treatment, the impact that the observer can have on the well-being of the stakeholder, by acting paternalistically, is smaller.

The experimental circumstances in treatment four and five are similar to the previous treatment. The only difference is that in treatments four and five, alternative B leads to a bonus payment of 20 EUR with a probability of 20% and 0% respectively.

3.1.4 Bonus options

In the experiment conducted by Altenau and Jensen (2017), there was no uncertainty on which bonus option would be in the best interest of the stakeholder. However, for an

interference to be paternalistic, it must be that it is done without the consent of the stakeholder (Dworkin, 2002; Grill, 2011). Therefore, I find it more realistic that a person that acts

paternalistically is not certain about the preferences of the person interfered with.

To create uncertainty on the stakeholder’s preferences, I included a lottery versus a certainty equivalent bonus option. In absolute terms, one of the two bonus options always result in a higher expected payoff. However, as the observer is unaware of the level of risk aversion of the stakeholder, he or she can never be certain which bonus option would be in the best interest of the stakeholder. This applies to all treatment levels, apart from treatment 5. In treatment 5, alternative A leads to a certain bonus payment of 5 EUR, and alternative B leads to a bonus payment of 0 EUR. Therefore, there is no uncertainty on which bonus option would be in the best interest of the stakeholder.

(23)

3.1.5 Sample

To conduct the experiment, in total 10 stakeholders and 101 observers were recruited. The choices made by the stakeholders served as an instrument for the second stage, to make sure the decisions made by the observers have real-life consequences.

In the second stage, in total 101 high school students participated and acted as observers. To ensure the quality of the responses, I had to exclude in total 3 observers from my analysis. All decided to interfere with the autonomy of the stakeholder in treatment 5 and, on his behalf, choose alternative B as bonus option. Hence, the observers are trying to decrease the well-being of the stakeholders, therefore, they are not acting paternalistically. Thus, the analysis the analysis of my experiment included the responses of 98 observers.

3.2 Hypothesis

In this section, I will give an overview of the main hypothesis of this experiment. In this experiment, I look at the willingness of people to make hard paternalistic interferences. In particular, I look at the average share of paternalists across the different treatments. This will allow me to examine the impact of these manipulations on the willingness to act

paternalistically.

Altenau and Jensen (2017) examined the causal impact of information asymmetry on the willingness to act paternalistic and found an increase in paternalistic interventions when the observer has an informational advantage. Following these findings, the first main hypothesis is as follows:

Hypothesis 1: The willingness to act paternalistically will increase when the observer has an informational advantage.

The second main hypothesis of the experiment is about the impact that the observer can have on the well-being of the stakeholder, by acting paternalistically. Research has shown that people value both the well-being and autonomy of others. Hence, they are confronted by a moral trade-off when making hard paternalistic decisions. As the impact on the stakeholder’s well-being increases, one would expect that people are more willing to act paternalistically. Hence, the second hypothesis is as follows:

(24)

Hypothesis 2: The willingness to act paternalistically will increase when the observer can make a larger impact on the well-being of the stakeholder.

Finally, Altenau and Jensen (2017) found that men are more likely to act paternalistically compared to women. This leads to the third hypothesis.

Hypothesis 3: Men are more likely to act paternalistically compared to women. An overview of the main hypotheses is presented in table 1.

3.3 Experimental variables

In this section, I will discuss my strategy to test the hypothesis and elaborate on the experimental variables.

In this experiment, there are five different treatments which vary in the level of information asymmetry and in the level of impact that the observer can have on the stakeholder’s well-being. The base treatment serves as a control treatment which will allow me to examine the impact of these manipulations on the willingness to act paternalistically.

The dependant variable is the willingness to act paternalistic. The willingness to act

paternalistically is measured as the share of observers who interfered with the autonomy of the stakeholders, by choosing a bonus payment on their behalf. For my analysis, I’m interested in whether the observers acted paternalistically or not.

In treatment 2 to 5, information asymmetry in favour of the observer is introduced. To test my first hypothesis, I will compare the means from treatment 2 to 5 with the base treatment to see whether they are statistically larger. If this is the case, then the willingness to act

(25)

The impact that the observer can make on the well-being of the stakeholder is the difference between the expected payoff of alternative A and B. Thus, the impact varies among the different treatments and is the largest in treatment 2 and 5, and the smallest in treatment 3 and 4. This will allow me to examine the sensitivity of the trade-off between how people value both the well-being and the autonomy of others. Table 2 below presents an overview of the stakeholder’s expected pay-off, and the impact that the observer can have on his well-being of each treatment.

To test my second hypothesis, I will analyse the means from treatment 2 to 5 to see whether they are statistically different from another. If the impact on the well-being of the stakeholder, which the observer can make, increases the willingness to act paternalistically, then the share of paternalists should be largest in treatment 2 and 5, and smallest in treatment 3 and 4.

Finally, I will compare the willingness to act paternalistically between men and women. First, I will compare the share of paternalists between men and women, irrelevant of the treatment level. Then, I will analyse the willingness to act paternalistically across treatments, by gender,

(26)

4. Results

In this section, the main findings of the experiment are presented. First, I will present the descriptive statistics, followed by the main results which will be used to answer the main research questions.

4.1 Descriptive statistics

The variable of interest is the share of paternalists for each treatment. Figure 1 presents an overview of the share of observers that acted paternalistically across treatments, referred to as paternalists.

In the base treatment, the observer and the stakeholder are equally informed on the bonus options. In this case, the stakeholder is in a better position to make the decision regarding his or her own pay-off, as the observer doesn’t know the stakeholder’s level of risk aversion. Yet, the share of paternalists in the base treatment is 35%. This indicates that 35% of the observers don’t place much value on another person’s autonomy, as they decided to interfere with the stakeholder’s autonomy while there is no moral trade-off.

(27)

It is also possible that some observers were unable to recognise that the stakeholder is in a better position to make the decision, as they didn’t take into account the stakeholder’s risk aversion. However, in this case, the observer and the stakeholder are still equally informed on the bonus options. Therefore, it seems that those observers that interfered in the base

treatment are confident that the stakeholder’s preference is inferior to their own preference. This indicates that, when acting paternalistically, over one third of the people prefer their own choices over the others’ choices.

In treatment two to five, the observers have an informational advantage. The share of paternalists is higher in all treatments, compared to the base treatment. Hence, the results indicate that people are more likely to act paternalistically when they have an informational advantage. Intuitively, it seems logical that the willingness to act paternalistically increases when the observers are better informed about the outcomes of the payment options. In the process of deciding whether to interfere or not, the observers will decide which bonus option is in the best interest of the stakeholder, from their personal perspective, and decide whether the stakeholder’s preference is inferior to their own. Being better informed on the matter means you can make a more educated decision. Since the observers are likely aware of this, I would expect that the willingness to act paternalistically is larger when the observers have an informational advantage.

In treatment two, the share of paternalists increases to 62%. This means that the observers are around 79% more likely to interfere with the stakeholder’s autonomy when they have an information advantage, and when they are in a better position to determine which payment option will result in the highest expected payoff. However, 38% of the observers did not act paternalistically. The observers were informed that they had an information advantage, and still more than a third of the observers chose not to interfere. Yet, as the majority choose to interfere, the results indicate that the majority values another’s well-being above his or her autonomy.

In treatment three, the impact that the observer can make on the stakeholder’s well-being is smaller compared to treatment two. Yet, the share of paternalists is equal for both treatments. Therefore, it seems that the level of impact that the observer can make on the stakeholder’s well-being doesn’t have an effect on the willingness to act paternalistically.

(28)

However, when moving to treatment four and five, the share of paternalists increases to 79% and 91%, respectively. Initially, I expected the share of paternalists to be equal in treatment three and four and in treatment two and five, as the impact that the observer can make on the stakeholder’s well-being is equal. However, the results do not follow these predictions.

Finally, to analyse whether the willingness to act paternalistically differs between men and women, I’ve compared the result by gender. Figure 2 presents an overview of the share of paternalists across treatments, by gender.

The results are similar to Figure 1, the share of paternalists is higher in all treatments, compared to the base treatment. The share of paternalists is equal in treatment two and three and increases as we move towards treatment four and five. However, irrelevant of the treatment level, the share of paternalists is larger among men, compared to women. This indicates that men are more willing to interfere with another person’s autonomy and that the moral preference for autonomy is smaller amongst men, compared to women.

It is also possible that men are more confident that their personal preference will be in the best interest of the stakeholder, compared to women. Yet, over 50% of the men chose to act

paternalistically in the base treatment, while there is no apparent reason for them to think that their preference is inferior to the stakeholder’s preference. A possible explanation is that men are less capable of recognising that they are unaware of the stakeholder’s risk aversion, compared to women. However, further research is needed to explain the reasons for this difference in paternalistic behaviour across gender.

(29)

4.2 Main results

In this section, the main hypotheses are tested. Furthermore, the results from the heterogeneity analysis are presented, to test whether the results are similar for men and women.

Result 1: There is a significant increase in the willingness to act paternalistically when the observer has an informational advantage.

To test whether the willingness to act paternalistically increases when the observer has an information advantage, I’ve conducted paired-samples t-tests between the base treatment and treatment two to five. The paired differences are presented in Table 3.

In treatment two to five, the observers have an informational advantage. There is a significant difference in the share of paternalists across treatment two to five, compared to the base treatment. These results suggest that the willingness to act paternalistically increases when the observer is in the position to improve the stakeholder’s well-being. This is in line with

Hypothesis 1, as more observers are willing to interfere with the stakeholder’s autonomy, when they have an informational advantage

Result 2: The impact that the observer can make on the well-being of the stakeholder has no apparent effect on the willingness to act paternalistically.

To test whether the willingness to act paternalistically depends on the impact that the observer can make on the well-being of the stakeholder, I’ve conducted paired-samples t-tests across treatments. The paired differences are presented in Table 4.

The impact that the observer can make on the well-being of the stakeholder is the difference between the expected payoff of alternative A and B. The impact on the stakeholder’s

(30)

well-being is equal for treatment two and five, and for treatment three and four. Therefore, I assumed that the share of paternalists would be equal between these treatment levels. In addition, I assumed that the share of paternalists would be larger in treatment two and five, compared to treatment three and four, as the impact on well-being is smaller for the latter.

First, there is a significant difference in the share of paternalists between treatment two and five, and between treatment three and four. Furthermore, treatment two is exactly equal to treatment three, and smaller than treatment four. Hence, the impact that the observer can make on the well-being of the stakeholder has no apparent effect on the willingness to act paternalistically.

To conclude, the results are not in line with Hypothesis 2. It seems that the level of impact on the stakeholder’s well-being only affects the willingness to act paternalistically when moving from treatment four and five. A possible explanation could be that the probability of actually receiving the bonus payment by the stakeholder influences the willingness to act

paternalistically. In treatment two and three, alternative B leads to the highest expected pay-off, where the stakeholder receives 20 EUR with a probability of 50% and 30%, respectively, while in treatment four and five, alternative A leads to the highest expected pay-off, where the stakeholder receives 5 EUR with a probability of 100%.

Result 3: The willingness to act paternalistically is significantly larger for men, compared to women.

In Table 5, the results of the logistic regression between the willingness to act paternalistically and gender is reported. Regression (1) shows that the willingness to act paternalistically is

(31)

smaller amongst women, compared to men, irrelevant of the treatment level. When taking a closer look at the different treatment levels, table 5 shows that women are less likely to act paternalistically in all treatment levels, except in treatment five.

In treatment five, the share of paternalists is very high, irrelevant of gender. A possible explanation could be that in treatment five, irrelevant of the level of risk averseness, alternative A always leads to the highest pay-off. Therefore, there is less uncertainty on the final outcome which could explain why both men and women tend to interfere with the stakeholder’s autonomy.

(32)

5. Conclusion

In this section, the main findings are summarized and discussed, and I will explain the main limitations of this study.

5.1 Main findings

This study analysed under what circumstances people are willing to interfere with another person’s autonomy, with the intent to promote his or her well-being. My approach was to study the sensitivity between two moral values when individuals are in the position to make hard paternalistic interferences in a non-hierarchical position, and how an informational advantage influenced the willingness to act paternalistically. When making hard paternalistic interferences people face a trade-off between these two moral values. Namely, another person’s autonomy and another person’s well-being. Whether a person interferes will reveal his moral preferences. What drives these moral decisions is still a complex field. This study aims to contribute to the understanding of this moral decision, and on the reason behind paternalistic behaviour when people are forced to make a decision between two moral values.

The main findings of this experiment are that there is a significant increase in the willingness to act paternalistically when the observer has an informational advantage, and that the

willingness to act paternalistically is significantly larger for men, compared to women. Also, I analysed whether the impact that a person can have on another person’s well-being, by acting paternalistically, positively affect the willingness to act paternalistically. I found no apparent effect. Finally, the results indicate that in general, people value another person’s well-being above his or her autonomy, as the majority of subjects decided to act paternalistically when they were in the position to improve the stakeholder’s well-being.

When the observer has an information advantage, he is significantly more likely to likely to act paternalistically. At least, 62% of the observers choose to act paternalistically when he or she had an informational advantage, which is 79% higher compared to the base treatment. Being better informed on the matter means you can make a more educated decision. Hence, acting paternalistically when you are better informed on the outcomes of a decision seems justifiable, as the majority of the observers choose to interfere in this case.

Furthermore, in treatment five, by acting paternalistically the observer was certain that he could improve the well-being of the stakeholder as only one bonus option would result in a

(33)

payoff for the stakeholder. Still, 9% of the observers choose not to interfere with the autonomy of the stakeholder. This indicates that some people do have an aversion against interfering with another person’s autonomy and value another’s autonomy over their well-being. Therefore, I assume that there does exist a moral trade-off between another person’s well-being and his or her autonomy.

Finally, I analysed the willingness to act paternalistically by gender, and men are statistically more likely to interfere with another person’s autonomy across all treatments, apart from treatment five. Therefore, adding uncertainty has a larger effect on the willingness to act paternalistically for women, compared to men. It is also possible that men are more confident that their personal preference will be in the best interest of the stakeholder, or that men are less capable of recognising that they are unaware of the stakeholder’s risk aversion, compared to women. However, further research is needed to explain the reasons for this difference in paternalistic behaviour across gender.

5.2 Limitations & future research

There are several limitations to this study. First of all, the number of participants is limited and did not include all members of the population. All participants were of a similar level of education, between 16 and 18 years old, and from Amsterdam. Therefore, external validity is low. Unfortunately, due to financial constraints, it was not possible to increase the number of participants, and to find a more representative sample.

Furthermore, there is a possibility that the participants didn’t fully understand the concept of paternalism. I excluded three subjects from my sample as they deliberately tried to minimize the payoff of the stakeholder by interfering with his autonomy and choosing the bonus option that would result in no payoff.

The findings of this study do imply that there exists a moral trade-off. Furthermore, I

examined the sensitivity of this moral trade-off between autonomy and wellbeing. Although I did not find an apparent effect between these moral values, this experiment analysed the moral trade-off in absolute terms. Therefore, further research is needed to examine how people value another’s autonomy and his or her well-being, taking into account the level of uncertainty.

(34)

References

Aghion, P., & Tirole, J. (1997). Formal and real authority in organizations. Journal of political economy, 105(1), 1-29.

Altenau, J. R., & Jensen, S. H. (2017). A study of moral motivation in paternalistic behavior: a study of how two variations of information asymmetry affect the willingness to make paternalistic decisions (Master's thesis).

Bartling, B., Fehr, E., & Herz, H. (2014). The intrinsic value of decision rights. Econometrica, 82(6), 2005-2039.

Berlin, I. (1969). Two concepts of liberty. Berlin, I, 118, 172.

Charness, G., Gneezy, U., & Kuhn, M. A. (2012). Experimental methods: Between-subject and within-subject design. Journal of Economic Behavior & Organization, 81(1), 1-8. Charness, G., & Rabin, M. (2002). Understanding social preferences with simple tests. The Quarterly Journal of Economics, 117(3), 817-869.

Christman, J. (2008). Autonomy in moral and political philosophy. Stanford encyclopedia of philosophy.

Choi, J.J., D. Laibson, B. Madrian, and A. Metrick. 2004. For Better or for Worse: Default Effects and 401(k) Savings Behavior. In: Perspectives in the Economics of Aging, Wise D (Ed.). Chicago: University of Chicago Press: 81-121.

DellaVigna, S. (2009). Psychology and economics: Evidence from the field. Journal of Economic literature, 47(2), 315-72.

Dworkin, G. (1972). Paternalism. the Monist, 64-84.

Dworkin, G. (2002, Nov 06). Paternalism. Stanford Encyclopedia of Philosophy: https://plato.stanford.edu/entries/paternalism/. First published Nov 6, 2002, substantive revision Feb 12, 2017 (retrieved 16. Sep.2017).

Feinberg, J. (1989). The moral limits of the criminal law: volume 3: harm to self. Oxford University Press, USA.

Feinberg, J. (1990). Harmless wrongdoing (Vol. 4). Oxford University Press. Fehr, E., & Camerer, C. F. (2002). Measuring social norms and preferences using experimental games. Working paper/Institute for Empirical Research in Economics, 97. Fehr, E., Herz, H., & Wilkening, T. (2013). The lure of authority: Motivation and incentive effects of power. American Economic Review, 103(4), 1325-59.

Frey, B. S., Benz, M., & Stutzer, A. (2004). Introducing procedural utility: Not only what, but also how matters. Journal of Institutional and Theoretical Economics JITE, 160(3), 377-401.

(35)

Grill, K. (2011). Paternalism. In Chadwick,R., Encyclopedia of Applied Ethics (2nd ed) Elsevier, http://kallegrill.se/texts/Paternalism%20preprint.pdf

Johnson, E.J. and D. Goldstein. 2003. Do Defaults Save Lives? Science 302, 1338-1339. Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1986). Fairness and the assumptions of economics. Journal of business, S285-S300.

Le Grand, J., & New, B. (2015). Government paternalism: Nanny state or helpful friend?. Princeton University Press.

Simon, H. A. (1951). A formal theory of the employment relationship. Econometrica: Journal of the Econometric Society, 293-305.

Simon, H. A. (1993). Altruism and economics. The American Economic Review, 83(2), 156-161.

Thaler, R. H., & Sunstein, C. R. (2008). Nudge- Improving decisions about health, wealth and happiness. London: Yale University Press. 

Thaler, R.H. and C.R. Sunstein. 2003. Libertarian Paternalism. American Economic Review Papers and Proceedings 93, 175-179.

Thaler, R.H. 2015. Misbehaving: The Making of Behavioral Economics. New York: W. W. Norton & Company

Thaler, R. H., & Benartzi, S. (2004). Save more tomorrow™: Using behavioral economics to increase employee saving. Journal of political Economy, 112(S1), S164-S187.

VanDeVeer, Donald, 1986, Paternalistic Intervention: The Moral Bounds on Benevolence, Princeton, NJ: Princeton University Press.

(36)

Appendix

1. Survey stage 1, including the instruction, round 1 and 2 (out of 10), and bonus questions.

Experiment

Welkom, in dit experiment wordt u gevraagd om 10 vragen te beantwoorden. Bij iedere vraag krijgt u een foto te zien van een nationale vlag met hieronder vier mogelijke landen. De vlag die u te zien krijgt hoort bij een van deze landen. Het is uw taak om de het juiste land te omcirkelen.

Voor uw deelname ontvangt u een vast bedrag van €5,-. Hiernaast krijgt u na het

beantwoorden van deze vragen de mogelijkheid om tussen twee alternatieven voor een bonus betaling te kiezen. Alternatief A en B.

Echter, er is een derde persoon die de mogelijkheid krijgt om deze keuze voor u te maken, of niet. Hoeveel geld u uiteindelijk verdiend aan uw deelname hangt dus af van uw eigen keuze, en die van een derde persoon.

Vul bij iedere ronde uw persoonlijke ID-nummer in. Uw ID-nummer is gelijk aan uw

telefoonnummer. Nadat het experiment is afgerond neem ik contact met u op om de betaling af te ronden.

(37)

ID-Nummer: ………..

Ronde 1

Van welk land is dit

de nationale vlag? Omcirkel 1 van de volgende opties a) Argentinië b) Kroatië c) Brazilië d) Colombia ID-Nummer: ……….. Ronde 2

Van welk land is dit de nationale vlag? Omcirkel 1 van de volgende opties a) Nepal

b) Mongolië c) Costa Rica d) Indonesië

(38)

Bonus Betaling (1/2)

Bedankt voor uw deelname. Voor uw deelname ontvangt u een vast bedrag van €5,-. Hiernaast krijgt u na het beantwoorden van deze vragen de mogelijkheid om tussen twee alternatieven voor een bonus betaling te kiezen. Alternatief A en B. Omcirkel uw keuze. Alternatief A: €5,- met 100% zekerheid

Alternatief B: €20,- met 50% zekerheid

Bonus Betaling (1/2)

Bedankt voor uw deelname. Voor uw deelname ontvangt u een vast bedrag van €5,-. Hiernaast krijgt u na het beantwoorden van deze vragen de mogelijkheid om tussen twee alternatieven voor een bonus betaling te kiezen. Alternatief A en B. Als u voor alternatief B kiest, dan doet u mee aan een loterij. De kans dat u €20,- ontvangt is onbekend.

Alternatief A: €5,- met 100% zekerheid

(39)

2. Survey stage 2, including the instructions of round 1 and 2 (out of 5)

Ronde 1

In dit experiment wordt u gevraagd om een keuze te maken die real-life gevolgen kunnen hebben voor een ander persoon.

Vorige week heeft een persoon een opdracht uitgevoerd. Dit is persoon X. Persoon X moest 10 rekenvragen beantwoorden. Door deel te nemen ontving hij/zij €5,-. Daarnaast komt persoon X ook in aanmerking voor een bonusbetaling. De bonusbetaling bestond uit de volgende twee alternatieven en persoon X mocht er één kiezen.

Bonus optie A: €5,- met 100% zekerheid Bonus optie B: €20,- met 50% zekerheid In deze ronde heeft persoon X de percentages gekregen

Na het maken van zijn/haar keuze heeft persoon X te horen gekregen dat een ander persoon kan beslissen of persoon X zelf de keuze moet maken tussen de betalingsopties, of dat deze andere persoon de keuze voor persoon X moet nemen.

U bent de andere persoon en kunt beslissen of de persoon zelf moet kiezen of u kunt namens hem of haar de betalingsopties kiezen. Uw beslissing is anoniem.

Vul uw antwoord op het antwoordenformulier in

Kies uit de volgende opties.

1. Ik kies ervoor dat persoon X zijn/haar eigen keuze maakt over welke betalingsoptie hij/zij zal ontvangen.

2. Ik kies ervoor dat persoon X niet voor zichzelf hoeft te beslissen en maak de keuze voor hem/haar

a. Ik kies ervoor dat persoon X optie A ontvangt b. Ik kies ervoor dat persoon X optie B ontvangt

(40)

Ronde 2