Current welfare measures and participation income: involvement
in socially valuable activities and motivational issues.
Abstract: This paper found that participation income can be considered a better policy than the current welfare system due to its ability to incentivize people to participate in socially valuable activities. First, this research argues participation income (PI) is the best alternative welfare measure available. Then, previous arguments made against the feasibility of PI are addressed through new technologies such as blockchain and new market development as the sharing economy. The new challenge identified for participation income is the definition of the right level of monetary compensation. To provide a quantitative answer to what the right level of compensation could be, a vignette survey was put in place. Results gave a significant negative effect of low monetary compensation level on individual’s willingness to participate. In the case of a high compensation, results are not significant. This can probably be explained by setting the level of the high compensation too low. Overall, this study pointed out that PI is a more sensible public policy than current welfare systems. It also contributed to stimulate and start a new discussion around a possible PI scheme.
Rebecca Belochi - 11084499 Bsc Economics and Business Finance and Organizations
University of Amsterdam Dhr. Dr. D.F. Damsma
Statement of originality
This document is written by student Rebecca Belochi, who declares to take full responsibility for the contents of this document.
I declare that the text and 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.
Table of Contents
1.
Introduction ... 5
2.
Ethical and political aspects on the idea of Participation Income ... 7
2.1. The dilemma of Participation Income, an ethically and politically superior idea considered unfeasible. ... 7
2.2. Emergence of new trends offering viable solution to PI’s dilemma ... 9
3.
Economic aspect on the idea of Participation Income: incentives and motivational
issues…. ... 10
3.1. The debate in academic literature surrounding motivation ... 11
3.2. Understanding of the relationship between intrinsic and extrinsic motivation, and monetary incentives ... 12
4.
Methodology ... 14
4.1 Research design ... 14
4.2 Vignette layout ... 15
4.3. Scales and measurements... 17
4.3.1. Scales ... 17
4.3.2. Control variables ... 17
4.3.3. Data manipulation ... 18
4.4. Data collection procedure. ... 18
4.5. Data analysis process. ... 19
5.
Results. ... 19
5.1. Descriptive Statistics ... 20
5.1.1. Correlations ... 20
5.2. Testing of hypothesis ... 21
5.2.1. Predictors. ... 21
5.2.2. Direct effect of low monetary compensation on willingness to participate... 22
5.2.3. Direct effect of high monetary compensation on willingness to participate ... 22
5.2.4. Effect sizes of extrinsic and intrinsic motivation in low monetary compensation setting. ... 22
5.2.5. Effect sizes of extrinsic and intrinsic motivation in high monetary compensation setting ... 23
5.2.6. Mediation effects in a low monetary compensation setting.... 23
5.2.7. Mediation effects in a high monetary compensation setting ... 23
6.
Discussion ... 26
6.1. Results summary ... 26
6.1.1. Low monetary setting ... 26
6.1.2. High monetary setting ... 26
6.2. Discussion of the results and perspective on previous literature findings ... 27
6.2.1. Crowding out of intrinsic motivation. ... 27
6.2.2. Higher incentives, higher performance ... 27
6.3. Contributions to theory and suggestions for future research ... 28
6.4. Contribution to policy makers ... 29
6.5. Limitations. ... 29
References ... 32
Appendix ... 37
Appendix 1A: no monetary compensation vignette ... 37
Appendix 1B: low monetary compensation vignette ... 38
Appendix 1C: high monetary compensation vignette ... 39
Appendix 2: Survey questionnaire ... 40
Appendix 3: overview of scales and original studies ... 43
Tables overview. Table 1: correlations overview for all variables... 20
Table 2: output summary of mediation analysis in a low monetary compensation setting. ... 24
Table 3: summary of direct and indirect effects of mediation analysis for a low monetary compensation setting. ... 24
Table 4: output summary of mediation analysis in a high monetary compensation setting. ... 25
Table 5: summary of direct and indirect effects of mediation analysis for a high monetary compensation setting ... 25
1. Introduction.
Current welfare systems are outdated. Our societies evolved and exposed individuals to new dilemmas. In the 1930s, John Maynard Keynes predicted that per capita income level would rise, allowing individuals to spend less time working to cover their basic needs and dedicate more time to leisure (1933). However, Keynes was only partially right, income per capita did increase but working hours remained unchanged. Robert and Edward Skidelsky (2012) argue one of the reason lies in individuals’ insatiable nature. Another reason could also be the way society frames work. As Keynes pledges in Economics possibilities for our grandchildren (1930), we need to evolve towards other systems enabling individuals to find a sense of purpose in other activities than a traditional work setting. Beyond this idealistic viewpoint, more grounded arguments also support the need for a new organisation of societies. The growing applications of artificial intelligence and robotisation are expected to have significant political and economic effects on societies all over the world. West (2018), underlines the political turbulences inherent to the disruption caused by artificial intelligence, especially when thinking of its impact on working activities. A McKinsey report published in 2017 expects the automatisation of jobs to impact 375 million workers worldwide. Similarly, Frey and Osborne (2013) identified more than 54% of European Union’s jobs as being at risk of computerisation. Therefore, as applications of artificial intelligence and robotisation increase, so does the range of individuals it might affect. Consequently, it is reasonable to argue the State needs to actively reshape how societies are organised in order to face the consequences of the development of such groundbreaking technology. The State has to offer opportunities to individual to redirect their interests. An area mostly controlled by the State, such as welfare, represents a good starting point. In that matter, the role of the State has to evolve and cannot be restricted to health and workfare. The economist Atkinson already developed a theory along this line of thought in 1996, Participation Income (PI). Individuals are monetarily compensated in return for their participation in the overall activity of the country. Atkinson’s definition of the participation requirement is broad, it can be interpreted as any activity that is benefiting to the society as a whole such as “ people working as an employee, engaging in approved forms of education or training, caring for young, elderly or disabled, and undertaking forms of voluntary work” (p.68-69). This paper investigates whether Participation Income is the key to reforming welfare systems and if this change is preferable to the current situation in terms of individual’s involvement in socially valuable activities.
The majority of the academic attention given to PI has been in comparison to Universal Basic Income (UBI). UBI differs from PI in its unconditionality, justified by the inherent right of each individual to freedom (Van Parijs, 2001). Until now, UBI was preferred to PI due to the simplicity of its operationalisation. All individuals are entitled to receive UBI, which does not constrain individuals’ right to privacy in contrary to Participation Income where increased monitoring and administrative overload burden the elaboration of a sensible scheme (Wispelaere and Stirton, 2007). However, the
making PI relevant again. Blockchain is introduced as the discovery of the century (Higginson et al., 2017). It consists of an open and decentralized database for transactions or information that are safely encrypted and impossible to hack. Trust in this peer to peer interaction system without a third party or official institution is build upon the infallible security system. Thus, Blockchain has the potential to revolutionize many sectors of our society, in the case of the public sector it could efficiently reorganize the way governments manage data (Cheng et al., 2017) while also enabling a better fraud prevention and more transparency (Higginson et al., 2017). Its potential is best illustrated through the Estonian “e-governance and digital society” e-Estonia initiative. During this process, blockchain was an essential component to allow trust in the new system and its operationalization. Citizens can, with their digital identity, access a variety of services from tax declarations, to health files and even vote (e-Estonia). This paper will not focus on the specificities of how the blockchain technology can support the creation of a PI scheme by diminishing the practical problems associated to it. Nevertheless, it relies on that argument to support the claim that PI is a viable solution for the public sector to consider today.
Alongside the fact that PI is relevant for society today, it is still necessary to determine whether or not it is preferable to the current welfare situation. If one follows the standard neoclassical economic reasoning, monetary incentives are expected to always have a positive impact on individual’s motivation, no matter the level of compensation. On the other hand, research in the field of economics and psychology found that monetary compensation does not always result in the increase of individual’s motivation and performance (Gneezy and Rustichini, 2000). In certain cases, a monetary incentive scheme can lead to worse results than one without any. This demonstrates the need for further research, in the context of participation income, to be able to determine whether its implementation would be beneficial for society or if current welfare systems are preferable.
The topic of Participation Income is broad. Current research is mainly qualitative and focused on PI’s opposition to UBI. For this paper, the aim is to quantitatively take part in the discussion surrounding PI in the hope to possibly contribute to the elaboration of a PI scheme. Whilst the participation criterion of Atkinson is the type of broad definition necessary for the implementation of PI, this paper will focus on socially benefitting activities that are not recognized by the State today. This choice was made due to the fact that activities such as education programs or full time employment are already part of public administrative structures. Therefore, it is expected to be easier to account for them and should not represent a challenge for a future PI scheme. Moreover, the additional value of PI rests in its ability to allow States to include until now unaccounted for, non-monetized, activities into the economic activity. Those can take the form of voluntary work, taking care of elderly, etc. In this sense, the goal is also to underline the importance of PI in enabling the State to take into account new types of activities. Accordingly, this paper will attempt to answer the following question: Are individuals more willing to engage in socially desirable activities not recognized by the State today in a Participation Income setting or in the current welfare system? Is this willingness dependent on the level of compensation?
The paper will be structured as follows. First, an outline of the political and ethical questions evolving around participation income is provided as well as an argumentation regarding the feasibility of a participation income scheme. Thereafter, the key literature around the economic aspect of PI, namely the opposition of intrinsic and extrinsic motivation, is discussed and leads to the elaboration of this research’s hypothesises. Then, the research methodology is addressed. Finally, after the results are analysed, they are put in perspective through a discussion and used to draw the conclusion.
2. Ethical and political aspects on the idea of Participation Income.
The concept of Participation Income is the result of Atkinson’s work to understand income inequalities and his attempt to introduce realistic proposals to significantly reduce them (2015, p.304). Current welfare systems in Europe fail to do so, mainly because of their dependence upon means-tested benefits (Atkinson, 1996). Mean-tested benefits are the monetary compensations allocated by the state to an individual or household under the condition that their “means”, i.e. level of income, is below a certain level considered critical. This system induces the perverse incentive for other individuals in a household to limit their employment to remain under the ceiling defined by policy makers that entitles them to support. Following Atkinson’s argument, new welfare proposal have to be developed. Participation Income provides individual monetary support independently of one’s working or familial situation and level of income under the only condition of participation to activities considered beneficial for society. The participation criterion is much more inclusive than the usual workfare requirement, it includes socially valuable activities such as “engaging in approved forms of education or training, caring for young, elderly or disabled, and undertaking forms of voluntary work…” (Atkinson, 1996, p.3) as well as any other type of valuable activity no recognised through monetisation.
2.1. The dilemma of Participation Income, an ethically and politically superior idea considered unfeasible.
Participation Income is not the only existing proposal with the intent of reforming the welfare system. The main opposition to PI is formulated as the Universal Basic Income idea. UBI is an universal compensation for all individuals, differentiated from PI by its unconditionality (Van Parijs, 2003). UBI’s main advocate, Philippe Van Parijs, supports the need for an unconditional measure through the external asset argument. Essentially, the argument rests on an egalitarian conception of justice and the individual right to freedom; all individuals should be entitled to receive a proportion of the available
admirable for its purist interpretation of justice and equality, PI stands out as a stronger proposal both ethically and in terms of ability to gather political support. First, academics and the public alike object to UBI principles out of fear of free riding, making it politically weak. For example, a referendum conducted in 2016 exploring public’s opinion on a potential basic income introduction in Switzerland faced substantial opposition with 77% of citizens voting against the project (BBC, 2016). The opposition can be understood through individual’s fear that a disconnect between money and work will be harmful for society. Therefore, as Wispelaere and Stirton (2007) point out, PI would allow to more significantly gather political support. It enables to reach common grounds between on one hand workfare supporters that require a participation criterion and on the other hand universalists who can come to terms with a broad and largely inclusive criterion. Additionally, PI also stands out as the most ethical proposal when confronted to UBI. Academics have also debated the risk of a too loose distribution of money associated with UBI. Brian Barry (2001), depicts this issue by analogy to a surfer in Malibu. In the context of a universal basic income, society suspects that individuals receiving benefits might use them to go surf on the beach instead of contributing in a meaningful way to society. This argument rests on the reciprocity principle introduced by Stuart White. It is defined as: “Those who willingly enjoy the economic benefits of social cooperation have a corresponding obligation to make a productive contribution, if they are so able, to the cooperative community which provides these benefits” (White, 1997, p. 317). However, as White underlines, the concept should not to be understood in a capitalistic way, contributions do not have to be equal to the benefits received or of the same nature. In this sense, a monetary compensation can be considered equivalent to a service, such as charity work or help to elderly. This interpretation of White’s principal coincides with the goal of Participation Income, incentivise individual to engage in socially valuable activities and offer an approach considered fairer or more ethical as to how individuals are eligible to receive monetary compensation.
However, Participation Income’s advantages also come with potential drawbacks, two are discussed here because they are considered to be the most substantial opposition to a PI scheme: administrative overload and the risk of a too government steered definition of socially valuable activities. The main point is addressed by Wispelaere and Stirton (2007), the risk of administrative overload. PI’s administrative weakness derives from the broad definition of the participation criterion which makes it difficult to enforce effectively and to monitor all individuals accordingly. The significant costs imposed on the public administration confront PI supporters with a dilemma between reducing the extent of the participation criterion to make it more easily enforceable, accept a weak enforcement of the broad criterion, or impose the high costs on the administration said justified by PI advantages for society. This challenge is introduced as the Trilemma of Participation Income (Wispelaere and Stirton, 2007). This administrative weakness consequently leads to political weakness since, in this situation, PI would not be the key to a political consensus. Additionally, another point of critique of Participation Income is that, in practice, the socially desirable activities citizens can engage in risk to be restricted to a centralised government steered definition. This would imply that the liberties
of individuals to engage in activities they consider meaningful are limited. This is embodied by the idea of Cristian Perez-Munoz (2015), a civic program designed by the State to help fulfil unmet social needs. Such a program can become very restrictive and feel like a duty more than a meaningful participation for individuals.
2.2. Emergence of new trends offering viable solution to PI’s dilemma.
The discussed drawbacks of participation income can be overcome today thank to the innovations and developments that took place in recent years, and therefore can be considered outdated. With regards to administrative overload, the blockchain technology represents a breakthrough to rethink public policies, especially welfare. Moreover, facing the risk of an oppressing presence of the State in defining socially beneficial activities, the recent developments of sharing platforms represent a viable solution to allow a decentralisation of activities.
The solution to administrative overload and intrusion lies in the blockchain technology. It’s applications are vast and, in the case of this research, very relevant for new public policies. For instance, Estonia has proven that the blockchain technology can digitalise and administratively operate an entire government (Arets, 2017). In fact, every Estonian citizen has a digital identity with which he or she can perform all administrative procedures online such as tax declaration, vote or access health files. Blockchain was a key element for this digitalised society to protect all the information processed and register all activities in a non-intrusive manner. Moreover, the procedure allowed to significantly reduce public expenditures by approximatively 2 percent of Estonia’s gross domestic product (Epstein and Franklin, 2017). This compelling decrease in administrative costs allows to draw the conclusion that, in the case of Participation Income, blockchain outdates the argument of its high administrative burden and would represent a considerable improvement compared to current welfare systems.
The second main point of critique against PI lies in the risk of a too restrictive government steered definition of socially valuable activities. In fact, the goal of Participation Income is to incentivise individuals to do more for the community, engage in socially beneficial activities. However, if the government restricts the definition of what a socially beneficial activity may be, thus reducing the choice of individuals, it takes the risk that they will engage in the opposite behavior. However, a growing trend over the last years has the potential to tackle this issue, the sharing economy. Sharing platforms demonstrated their potential to efficiently match local demands and supply thanks to the internet. It great potential value is best illustrated through the estimated 23 billion dollars of funding directed towards sharing platforms between 2010 and 2017 (Wallenstein and Shelat, 2017). Benkler (2004) defines the sharing economy as a potentially more efficient third organising entity of economic production than the State and markets.
To conclude, as a result of the facts given above the discussion around PI’s implementation reopens. New innovations enable PI to be regarded as a relevant and legitimate alternative welfare measure. Today, a viable PI scheme could take the form of a sharing platform, allowing individuals to participate in a way that is significant to them, operated by blockchain for more efficiency, security and independence. Howbeit, as participation income’s feasibility was the center of academic debates and research in the last years, a critical aspect of its scheme-design has been neglected. Indeed, the ability of PI to live up to its ideal essentially depends on the level of compensation. Neoclassic economic theory suggests the only way to incentivise individuals to exert an effort is a monetary compensation (Gneezy and Rustichini, 2000). It follows that any monetary compensation level should have a positive effect on individuals’ willingness to engage in socially valuable activities. However, research in behavioural economy and psychology has proven otherwise. In fact, monetary compensation might negatively affect individuals’ willingness to exert an effort (Gneezy and Rustichini, 2000) i.e. in some cases, individuals perform better without a monetary reward. Additionally, the effect of the monetary incentive also depends on the initial motivation of the individual and the context. A person engaging in an activity for other reasons than make a living will see his or hers motivation decrease with the offer of a monetary reward (Heyman and Ariely, 2004). This opposition in the literature suggests the determination of whether or not a monetary compensation is necessary and what its level should be are critical to the success of a participation income scheme and thus this point is addressed in the next section.
3. Economic aspect on the idea of Participation Income: incentives and motivational issues. Now that the ethical and political character of PI have been discussed and the idea of a PI scheme made relevant again thanks to the development of today’s new technologies, it is important to investigate what the next interrogation might be when thinking of Participation Income. The new challenge that PI advocate face is the question of monetary compensation for individual’s participation. In fact, the possible benefits of a Participation Scheme in comparison to the current welfare system can only be maximised if the level of compensation is right. Literature in economics, social sciences, and psychology reveals a complex discussion surrounding the effects of compensation on individual’s motivation and performance.
3.1. The debate in academic literature surrounding motivation.
The first intuition on the role of intrinsic motivation was officially formulated by Titmuss (1970), he defended the idea that monetizing blood donations would decrease individual’s sense of duty and lead to a decrease in willingness to donate. However, until now, economics have been almost entirely relying on the relative price effect theory to explain changes in supply, production, motivation, performance… (Frey and Jegen, 2001). This is expressed through three different views on motivation in economics. One view sees an individual’s motivation as very hard to define and identify as intrinsic or extrinsic. In this case, intrinsic motivation is deliberately discarded and defined as an endogenous constant in economic models (Frey and Jegen, 2001). Another consideration is the one of the classical economists, they recognize the value of intrinsic motivation but also acknowledge the complexity of its inclusion in economic theories. In this line of thought, Adam Smith stressed the importance of morals and sense of duty for individual’s action and the market place and how extrinsic rewards can distort those values in Theory of Moral Sentiments (1822). Finally, another approach considers intrinsic motivation to be completely irrelevant. Lazear (2000), for example, discredits the value of recognizing intrinsic motivators whilst studying the effectiveness of pay-for-performance schemes in a monotonous work setting; placing windshields on cars (p.1342).
Moreover, not only the sources of motivation are discussed throughout academic fields but also the effects of different type of incentives on motivation. Ariely, Gneezy and Loewenstein (2009) point out that high monetary incentives can lead to worse performance. They reveal that large monetary incentives do not lead to higher performance in tasks requiring creativity, problem solving and concentration. Empirical evidence on the effect of incentives has since then grown. Deci, Koestner and Ryan (1999) define through the Cognitive Evaluation Theory the underlying motivators of intrinsic motivation: the need of an individual for autonomy and competence. This leads to the conclusion that monetary rewards have a negative effects on intrinsic motivation and, in a majority of cases, this is due to the perceived controlling effect of the incentive. Heyman and Ariely (2004) also underline the importance of the level of incentives. They show that low incentives crowd out intrinsic motivation and that, in the case of extrinsically motivated behavior, higher incentives are more effective. Those results have been taken in consideration in economics even though they cannot be completely explained. The crowding out effect of intrinsic motivation remains one of the most substantial anomalies in economics (Frey and Jegen, 2001). For the purpose of this research however, it is necessary to analyze the effect of different levels of incentives on motivation or intention to participate to be able to analyze Participation Income to its fullest potential. Hence, the new relevant challenge PI faces is how to frame monetary compensation allocated to individuals to avoid crowding out intrinsic motivation and still incentivize individuals to participate. Moreover, the question does not limit to whether or not compensation is necessary but also what the right level of compensation should be.
3.2. Understanding of the relationship between intrinsic and extrinsic motivation, and monetary incentives.
First, the relationship between low monetary reward, no compensation at all, and individual’s motivation is studied. Frey and Oberholzer-Gee (1997), linked the discussion on monetary incentives to public policies’ elaboration. They stressed the importance of taking into account the possible negative effect of monetary incentives in the case of public policies that target areas where involvement depends on individual’s intrinsic motives. Also, Gneezy, Meier and Rey-Biel (2011) raise the point that monetary incentives can be a source of misinterpretation, a signal, making the task of determining the right level of compensation very delicate. For instance, offering a monetary incentive could be understood as a lack of confidence in one’s intrinsic motivation to perform a task correctly. Additionally, Gneezy and Rustichini (2000) defend the view that money destroys individuals’ sense of duty and therefore often leads to less motivation and less willingness to exert an effort. Through an experimental study in day-care centers in Israel, they investigated the relationship between monetary incentives and behavior. A fine was appointed to parents that were late to pick up their children. Surprisingly, imposing fines did not decrease the amount of late parents each day but lead to the opposite behaviour. Parents perceived the fine as an opportunity to diminish the morals behind the duty to be on time and avoid being late, and simply replaced it by monetary value. This case illustrates how money can drive out intrinsic values. Another point they raise is the importance of the level of compensation in the case a monetary reward is offered. They have quantitatively examined whether or not higher incentives leads to better results and concluded that indeed, if a monetary compensation is offered, the higher the compensation the better the performance. This entails that a low monetary incentive would, in addition to having the possible drawback of crowding out intrinsic motivation, have a lesser effect on performance than higher incentives if the effect is expected to be positive. Therefore, it seems crucial to determine whether a small monetary compensation leads to better performance than no monetary compensation at all and how it affects intrinsic and extrinsic motivation.
Secondly, the relationship between a high monetary compensation, no compensation and an individual’s motivation is examined. Heyman and Ariely (2004) underlined that when increasing the monetary incentive, performance or motivation similarly increase. Moreover, Reeson and Tisdell (2008) have specifically studied the relationship of monetary incentives and individual’s contribution to public goods. Those contributions are mainly motivated by intrinsic factors and therefore are very sensitive to monetary incentives. They identify a dilemma for policy makers to strengthen one’s already existing engagement by intrinsic motivation while providing a good extrinsic incentive to encourage other participants. In this context, a low monetary compensation is not the solution since it would only crowd-out individual’s intrinsic motivation. However, a high compensation could enable to not adversely affect intrinsically motivated individuals whilst providing a good incentive for extrinsically inclined individuals. This would result in a higher contribution to public good overall, both intrinsically and
extrinsically motivated. Bowles and Polonia-Reyes (2010), identify another reason why in some cases monetary incentives can enhance a pro-social behaviour: situational cues or how incentives are framed. If a monetary incentive is perceived as fair (this entails a level of compensation that is not considered too low and perceived as disrespectful) and complementing one’s social preferences, there is a high chance it will lead to more willingness to participate. Therefore, the effect of a higher monetary compensation should be analysed in more depth to be able to identify its effect on both intrinsic and extrinsic motivation and the resulting effect on individual’s willingness to participate.
To conclude, as a result from the analysis of motivation literature and the relationship between intrinsic and extrinsic motivators, several points can be made. First, two possible situations stand out when thinking of the new challenge PI is facing: identifying the right level of monetary compensation. It is clear that the relationship between monetary incentives and willingness to participate is not as straightforward as it was initially assumed. This is mainly due to intrinsic and extrinsic motivation that react differently to incentive levels. In the case of PI, contribution is mostly dependent on intrinsic factors as the goal is to participate meaningfully to the country’s activity. However, extrinsic motivation can also be relevant to increase the amount of participation. Therefore, the two settings that need further investigation examine a low level of monetary compensation and a high level in relation to individual’s willingness to participate. They are the basis for the hypothesis tested in this research, defined as follows.
H1: A low monetary compensation setting will result in less willingness to participate than a setting with no monetary compensation.
H2: A high monetary compensation setting will result in more willingness to participate than a setting with no monetary compensation.
H3: Intrinsic motivation is a stronger driver of individual’s willingness to participate than extrinsic motivation in a low monetary incentive scheme compared to no monetary compensation.
H4: Extrinsic motivation is a stronger driver of individual’s willingness to participate than intrinsic motivation in a high monetary incentive scheme compared to no monetary compensation.
H5: The relationship between individuals’ willingness to participate in a low monetary setting compared to no compensation setting is sequentially mediated by extrinsic and intrinsic motivations, such that the presence of low compensation (vs. no compensation) is associated with higher (lower)
extrinsic motivation and reduced (higher) intrinsic motivation, which leads to lower (higher) intentions to participate.
H6: The relationship between individuals’ willingness to participate in a high monetary setting compared to no compensation setting is sequentially mediated by extrinsic and intrinsic motivations, such that the presence of high compensation (vs. no compensation) is associated with higher (lower) extrinsic motivation and reduced (higher) intrinsic motivation, which leads to higher (lower) intentions to participate.
As previous critics of Participation Income have been answered, PI established as an ethically and politically superior idea to other welfare measures, and the new relevant challenge for PI identified, the next step to take is the quantitative process of this research to provide an answer to the research question.
4. Methodology.
The methodology this paper follows is described here. First, the choice of research design is defended. Then, the set up of the research is clarified through the elaboration of the vignettes and scales. This is followed by a description of the data collection procedure. Finally, the data analysis process is laid-out.
4.1 Research design.
To offer a quantitative answer to the question this paper raises, a vignette survey design was chosen. Also known as factorial study, it consists of a vignette experiment associated with a standard survey questionnaire. The added value of such a research design is discussed here.
An experimental vignette study, as defined by Aguinis and Bradley (2014), “consists of presenting participants with carefully constructed and realistic scenarios to assess dependent variables including intentions, attitudes and behaviours”. More specifically, a vignette, “a short carefully constructed scenario” (Atzmuller and Steiner, p.128, 2010) is presented to participants before they answer a standard survey questionnaire. The added value of the vignette survey thus lies in the possibility for researchers to randomly present different scenarios to participants and control the causal relationship between variables. A vignette study therefore enables to simultaneously reinforce the internal and external validity of a study, which has been one of researcher’s most prominent dilemma when studying causal relationships so far (Aguinis and Bradley, 2014). Traditionally, the use of an experiment yields to high internal validity due to the controlled setting but low external validity since
findings cannot be generalised. On the other hand, a non-experimental setting such as a survey leads to a low internal validity due to the unclear relationship between variables whilst external validity remains good (Atzmuller and Steiner, 2010). In the case of a vignette experiment, internal validity is strong due to the controlled scenario participants are presented whilst it remains representative of reality and therefore strengthen external validity as well.
When thinking of Participation Income and the best investigation method to evaluate if it is a better welfare measure, two points stand out. First, it is essential to realise that participants need to face a realistic context to make sure they answers are representative so that internal validity is strong. A big scale real-life experiment would be very complex to set up and costly. Moreover, it is also important to be able to control precisely for certain variables when trying to measure intention to participate in a PI scheme, since as it was discussed before it can already be dependent on two factors: intrinsic and extrinsic motivation. The survey questionnaire added to the vignette experiment allow to present individuals a controlled realistic setting whilst remaining an accessible method and therefore is the best option for this research.
4.2 Vignette layout.
A realistic description of a possible participation scheme had to be elaborated to provide participants with a credible scenario. The construction of the vignette is essential since it is the key element to participant’s answer to the survey questionnaire. If the vignette is not believable, respondents’ answers will not be representative. Therefore, it is essential to clearly layout the details participants received.
For this research, three vignettes were created. Each one was developed with the objective to provide an answer to the developed hypothesis. Essentially, all three vignettes are completely similar except for the level of compensation individuals receive. This means that all three vignette present the same set of survey questions to the participants and the description of the PI scheme is also similar. For the first vignette, no monetary compensation is offered. Individuals engage in socially valuable activities because they enjoy helping others. For the second and third scenarios, individuals are respectively offered a low and a high monetary compensation. Consequently, this setting allows to control the dependent variable: intention to participate, given the level of compensation by randomly assigning participants to one out of the three vignettes. The decision on the level of low and high monetary compensation was made based on the level of the minimal hourly wage in the Netherlands, which is 9.96 euros an hour for an individual working full time (Rijksoverheid, 2018). In the high monetary setting, a participants is compensated at the value of 10 euros an hour, free of tax. Therefore, if he decides to dedicate the equivalent amount of time of a full time job, he shall be able to subsist to
participants can earn up to 1600 euros free of tax. In the low monetary compensation setting, individuals are offered 5 euros an hour. Here, even if individuals decide to invest all their time in the project, they will not receive enough money to subsist, as the maximum represents 800 euros a month, free of tax. In this case, PI is seen as a way to help others next to one’s primary activity to make a living.
The project described in all three vignettes is a government developed online platform and app independently operated by blockchain called “Live better together”. The idea is simple, individuals are put in contact to help each other out, providing or offering help at a local level. There are no restrictions as to which activities individuals wish to engage in. The community consists of individuals but also local governmental agencies and non-profit organisations. Any one of them can post requests for help but also answer them. Possible requests on the platform could take many forms. For example, it could be a neighbour needing help with administrative work, supervision and academic tutoring for his or her children, etc. But it could also be a post from the municipality asking help cleaning up the forest or a post from the shelter asking for help walking dogs. The goal of the platform is therefore to incentivise individuals to engage in many diverse socially beneficial activities by offering a wide range of them. Individuals can dedicate up to 40 hours from their time per week to socially valuable activities. Hence, it is important to note that this is a crucial aspect in how PI represents an alternative to current welfare systems. If one considers the high monetary compensation scenario, individuals by participating 40 hours a week can earn the equivalent of the minimum wage and therefore all individuals looking to work and participate in the country’s activity are able to do so. An overview of the three vignettes as well as the survey questionnaire can be found in appendices 1 and 2.
It is also important to note that blockchain holds a crucial role in the execution of all administrative work and supervision. The first step in which blockchain is essential is individual’s registration. To make the registration process as quick and non-intrusive as possible, citizens only have to enter basic information such as name, age, familial situation, address and the blockchain will cross-reference it with the data base of the government. Additionally, activities individuals which to engage in have to be described in a couple of keywords that will also enable the blockchain to analyse the goal intended with the activity and verify it is beneficial for society. Moreover, blockchain also has a central role to register individual’s activity, this is done through a double-rating system. Both parties engaged in an activity have to provide some references such as the date, the duration and their level of satisfaction; if the provided information match on both sides the blockchain registers the activity as completed and approved. Finally, blockchain is also key to the distribution of the monetary compensation offered by the government. This is done through the elaboration of smart contracts, self-initiating contracts, that are registered with the blockchain. In the case of this project, payment is initiated if the activity registered by the blockchain on the platform meets the requirements set by the government.
4.3. Scales and measurements.
As it has been previously established, individual’s intention to participate depends on more than one factor. Indeed, extrinsic and intrinsic motivation seem to react differently to levels of monetary incentives from previous academic researches. Moreover, one could also reasonably assume that an individual’s intention to participate is dependent on age, working situation, etc. The scales and different measurements used in this research in an attempt to clarify which factors significantly affect individual’s willingness to participate are described here.
4.3.1. Scales
To insure that the internal validity of the experiment is strong, all the scales were adapted from previous peer-reviewed and published researches targeting the same variable. Additionally, to reinforce the reliability of the intended measurements, one scale always comprised at least three items. They were arranged in an specific order, to diminish the chance of participants realising two questions intended to measure the same intention. Finally, all questions were associated to a seven point scale when possible.
The scales used in this research include individual’s intention to participate, extrinsic motivation, and intrinsic motivation. Individual’s intention to participate was measured through scales developed by Pavlou and Gefen (2004) as well as White et al. (2012) and Hamari et al. (2015). Pavlou and Gefen investigated how to enhance individual’s trust in online market places and encourage online transactions. White et al. researched the link between one’s intention to purchase a product and its ethical nature. Finally, Hamari et al. investigated individuals’ motivation for engaging in collaborative consumption also known as sharing economy. All three studies developed questions on respectively transaction intentions, purchase intentions and intention to participate. Intrinsic motivation scales were developed by adapting questions from the research of Paul et al. (2009) which focused on the source of individual’s repeating purchase behaviour to improve customer relationship. Measurements for extrinsic motivation were taken from the research of Hamari et al. (2015). In this case, involvement in the sharing economy was analysed under the angle of monetary motives. The adapted questions for this study are exhibited in appendix 3.
4.3.2. Control variables
Other measurements relate to participants’ personal information, relevant for this research. They include participant’s age, working situation, residency, level of income, gender, education, as well as their involvement in volunteering activities. Bucher et al. (2016) determined previous involvement in the volunteering activities usually positively affects individual’s willingness to participate in the
participation in collaborative consumption. Additionally, Ranzini et al. (2017) showed that millennial, and thus younger citizens with lover income levels and less stable employment opportunities, are more sensible to the new trend of the sharing economy. Finally, residency is considered to be a relevant criterion for this research as participants range from different countries across the European Union and a pattern of support for a PI scheme could be identified depending on the existing national welfare policies.
4.3.3. Data manipulation
The data was first analysed in terms of complete data sets. All incomplete participations were deleted. In total, 120 complete responses were collected and downloaded in SPSS.
Afterwards, an investigation for outliers in the data set was conducted. Anomalies in the data were found in how individuals value the compensation they were offered. Five participants in the low compensation setting perceived their hourly compensation as high, whereas six participants in the high compensation setting saw their compensation as low. As individual’s perception of their compensation is a key element in predicting their intention to participate, since it is assumed it can affect both intrinsic and extrinsic motivation, those 11 participants were categorised as outliers and deleted from the data set.
4.4. Data collection procedure.
Here, the process of data collection is described. The goal was to gather as many answers as possible from diverse backgrounds. The main mean of distribution was social media. Respondents principally represented the Netherlands, Germany and France. Additionally, potential participants were approached individually and asked for their email addresses to complete the survey in public places; for the majority at the university of Amsterdam. The survey was developed in English. However, to make participation of French individuals easier, the survey was also translated and available in French. To make results statistically significant, a goal of about forty participants per scenario was fixed, corresponding to a total of 120 complete data sets.
4.5. Data analysis process.
Overall, the data set is composed of 62 female and 46 male participants, represented by a mean age of 34.9 and individuals ranging from 15 to 68 years old. The majority of them are divided amongst three nationalities, namely Dutch, French, and German representing respectively 56%, 21%, and 11% of the pool of nationalities. A majority of participants, 36%, are also students as opposed to individuals working full time, 27%, and self-employed 19%. The mean salary of the sample is situated somewhere between zero and 1,550 euros a month after tax, probably due to the high number of students in the sample. There is also a good distribution of higher incomes ranging above 3,500 euros a month, 30%, and incomes between 1,550 and 3,500 euros a month representing 16%. Finally, of all the participants, 35% claim to already have been involved in volunteering activities.
All non-continuous variables were transformed into dummy variables. This resulted in dummies for nationality, salary, gender, age, and previous involvement in volunteering activities (see appendix 4). Also, one new variable was created for each measurement namely intrinsic motivation, extrinsic motivation and intention to participate. The correlation of the items composing the new variable was checked, to make sure they all measure the same intention, through Cronbach’s alpha. This step was taken to ensure the internal validity of the study remains strong and measures what it intends to investigate.
Finally, once all these operations were executed, the following tests were applied to the data set. To begin, a correlation table is set up to allow comparison of all variables and their effects on each other. Afterwards, the relationship between intention to participate and the monetary compensation level is studied thanks to a PROCESS test. The Hayes PROCESS test is a regression based technique that enables to control for the mediating effect of certain variables (Hayes, 2011). In the context of this study, as it is attempted to show the effect of intrinsic and extrinsic motivation on intention to participate, the two-mediator model six is used to test both hypothesis.
5. Results.
In this section, the results of the analysis described in the previous section are presented. First, the correlation table is discussed as it clarifies the relationship between variables to allow for further investigation. Thereafter, the mediating effects of intrinsic and extrinsic motivation on intention to participate are presented.
5.1. Descriptive Statistics. 5.1.1. Correlations.
To begin with, a simple correlation analysis was used to identify whether the variables did have a basic linear relationship with one another, overall results can be found in table 1. Most importantly, extrinsic motivation (r = .38, p < .01) and intrinsic motivation (r = .77, p < .01) were both associated with higher intention to participate which is in line with the reviewed literature. This was a crucial point for the elaboration of a double mediation analysis since both variables are expected to mediate individual’s intention to participate. Both variables also seem to be connected since extrinsic motivation is positively associated to intrinsic motivation (r = .32, p < .01). Additionally, extrinsic motivation appears to be positively associated with a low monetary compensation setting (r = .23, p < .05) as well as with a high monetary compensation setting (r = .24, p < .05). However, intrinsic motivation does not show any significant result associated to either low (r = -.08, p = .389) or high (r = -.001, p = .989) compensation even though the implied negative relationship is in accordance with the literature. Thus, even though results are not significant, it is argued the relationship is still worth investigating further in a mediation setting.
Table 1: correlations overview for all variables.
Note: N=109. * p< .05; ** p<.01 1 2 3 4 5 6 7 8 9 10 11 12 1. LowCom 2. HighCom -.51** 3. Part_T -.16 .07 4. Intr_T -.08 .00 .77** 5. Ext_T .23* .24* .38** .32** 6. Age -.06 .12 .12 .04 -.12 7. French -.06 -.11 .01 .01 -.18 .08 8. Other_N -.04 .06 .12 .07 .03 .10 -.28** 9. INCOME_M -.02 .07 -.03 -.14 -.17 .15 .19 -.16 10. INCOME_H -.09 .09 .04 .16 -.15 .57** .17 .08 -.23* 11. FEMALE -.02 .00 -.01 .03 -.05 .12 -.07 .01 .08 .00 12. VOLUN_Y -.11 .24* .05 -.01 .01 .07 .03 .19* .15 -.04 .05
5.2. Testing of hypothesis.
All hypothesis are tested through Hayes’ serial multiple mediator model number six. The steps taken follow his recommendations as given in his book (Hayes, 2013, pp. 143-163). How results are reported and what paths are tested can be seen in figure 1. The model was run with the consistent heteroscedasticity consistent standard error option of Hubert and White, as recommended by Hayes and Cai (2007), to allow for a stronger validity and power of the tests performed. Moreover, the tests were run whilst also controlling for age, gender, nationality, and level of income. The results can be found in tables 2, 3, 4, and 5.
Figure 1: model 6, double mediation.
Note: taken from Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Publications.
5.2.1. Predictors.
When running the analysis, certain control variables were found as predictors of the two mediators and the dependent variables in both monetary settings. They will be presented here before addressing hypothesis testing.
Two control variables significantly associate to the dependent variable willingness to participate. First, it is found that age is positively associated to intention to participate (B = .013, SE = .004, p = .002). This implies that the older an individual is, the more likely he is to participate in the proposed participation income scheme. Additionally, high income appears to be negatively related to an
individual’s willingness to participate (B = -.358, SE = .151, p = .019). This can be interpreted as the idea that individual’s with higher incomes will be less willing to engage in a participation scheme.
Moreover, categories of incomes also seem to have effects on both mediators. First, high income is perceived as positively paired with intrinsic motivation (B = .497, SE = .209, p = .019). This can be understood as individuals earning higher incomes to be more sensitive to intrinsic motivation. Also, medium income negatively affects extrinsic motivation (B = -.860, SE = .337, p = .012). This implies that individuals earning between 1,550 and 3,500 euros a month are adversely affected by extrinsic motivation.
5.2.2. Direct effect of low monetary compensation on willingness to participate.
Hypothesis one predicted that a low monetary compensation setting will result in less intention to participate than a setting with no monetary compensation. The total effect cis negative and significant (c = -.413, SE = .200, p = .042) which means that a low monetary compensation setting versus no monetary compensation is negatively associated with individuals’ willingness to participate (see table 3). Therefore, hypothesis one is supported.
5.2.3. Direct effect of high monetary compensation on willingness to participate.
Hypothesis two speculated that a high monetary incentive scheme will result in more willingness to participate than no monetary incentives at all. This can be tested through the total effect of the compensation scheme on intention to participate. The total effect (c = -.099, SE = -.210, p = .639) is not significant (see table 5), thus hypothesis two rejected. There is no strong statistical support for the idea that a higher level of compensation results in more willingness to participate. However, the latest discussions on mediation methods (Hayes, 2009) argue that scholars should continue to test for mediation in the absence of direct effect.1 Thus, this paper will proceed with the mediation analysis. 5.2.4. Effect sizes of extrinsic and intrinsic motivation in low monetary compensation setting.
Hypothesis three assumed intrinsic motivation is a stronger driver of intention to participate than extrinsic motivation in a low monetary incentive setting. By comparing partially standardised indirect effects, one can see that indeed intrinsic motivation has a bigger effect (Ind2= -.502, SE = 175, CI: -.840 to -.157) on intention to participate than extrinsic motivation does (Ind1 = .251, SE = .112, CI: .075 to .499) as it can be seen in table three. Hence, hypothesis three is supported.
1 “ If a mediator is a variable, M, that is causally between X and Y and that accounts at least in part for the association
5.2.5. Effect sizes of extrinsic and intrinsic motivation in high monetary compensation setting. Hypothesis four tested that in the case of a high monetary compensation setting, extrinsic motivation would be a stronger driver of intention to participate than intrinsic motivation. Partially standardized indirect effects allow to provide an answer. However, extrinsic motivation appears to have less of an effect (Ind1 = .277, SE = .117, CI: .082 to .534) on willingness to participate than intrinsic motivation (Ind2 = -.460, SE = .180, CI: -.837 to -.117). Therefore, even though the magnitude of the negative effect of intrinsic motivate does decrease (from -.502 to -.460), as it can be seen by comparing tables three and five, it also remains larger than the effect of extrinsic motivation and accordingly, hypothesis four is rejected.
5.2.6. Mediation effects in a low monetary compensation setting.
Hypothesis five predicted that intrinsic and extrinsic motivation would sequentially mediate the relationship between a low monetary compensation and individual’s willingness to participate.
After running the analysis, the direct effect of low monetary compensation on willingness to participate remained significant (c’ = -.531, SE = -.145, p = .001) but the value of the total effect is smaller (c = -.444, SE = .200, p = .042). From this, we can infer there is an associated effect with at least one of the mediators.
The mediational path running from low monetary compensation, to intrinsic motivation until intention to participate is significant and negative (Ind2=-.502, SE = 175, CI: -.840 to -.157). Also, the mediational path from low monetary setting to extrinsic motivation on behavioural intentions is significant and positive (Ind1 = .251, SE = .112, CI: .075 to .499).
Finally, this study finds evidence of sequential mediation. In this case, extrinsic motivation positively mediates the path from intrinsic motivation to intention to participate (Ind3 = .338, SE = .108, CI: .153 to .570). However, against expectation this effect is positive and thus does not support hypothesis five which is rejected.
5.2.7. Mediation effects in a high monetary compensation setting.
Finally, hypothesis six predicted whether the relationship between a high monetary compensation and intention to participate is sequentially mediated by intrinsic and extrinsic motivation.
First, there is a clear mediating effect of intrinsic and extrinsic motivation on intention to participate. The direct effect when testing for all the path simultaneously is not significant (c’ = -.298, SE = .159, CI: -.594 to .039), this evidence supports the claim that intrinsic and extrinsic motivation fully mediate the relationship between X and Y.
Moreover, the mediational path from a high monetary compensation setting to intrinsic motivation until individual’s willingness to participate is significant and negative (Ind2 = -.460, SE =
motivation until intention to participate is positive and significant (Ind1 = .277, SE = .117, CI: .082 to .534).
Finally, this study proves sequential mediation. Whilst the mediational path testing for both intrinsic and extrinsic motivation is significant (Ind3 = .374, SE = .110, CI: .167 to .602) it is positive. Therefore, evidence goes against the hypothesis of this research that extrinsic motivation negatively relates to intrinsic motivation and hypothesis six is rejected.
Table 2: output summary of mediation analysis in a low monetary compensation setting.
Consequent
M1 (extrinsic) M2 (intrinsic) Y (intention to participate)
Antecedent Coeff. SE p Coeff. SE p Coeff. SE p
X (low vs no) 𝑎1 1.292 .243 p < .001 a2 -.668 .234 p = .005 c' -.494 .145 p = .001 M1 (extrinsic) - - - d21 .348 .072 p < .001 b1 .180 .053 p = .001
M2 (intrinsic) - - - b2 .699 .068 p < .001
constant iM1 3.138 .361 p < .001 iM2 4.675 .338 p < .001 iY .506 .155 p = .205 Control variables Age - - - .013 .004 p = .002 High Income - - - .497 .209 .019 -.358 .151 p = .019 Medium Income -.860 .337 p = .012 - - - - R2 = .295 R2 = .240 R2 = .684 F(8,96) = 7.936, p < .001. F(9,95) = 4.275, p < .001. F(10,94) =31.737, p < .001.
Table 3: summary of direct and indirect effects of mediation analysis for a low monetary compensation setting.
Effect Effect_ps* SE p LLCI ULCI
Direct effect c' -.494 -.531 -.145 p = .001 -.781 -.206 Total effect c -.413 -.444 .200 p = .042 -.811 -.015 BOOT SE BOOT LLCI BOOT ULCI Indirect effect 1_ps* a1,b1 .251 .112 .075 .499 Indirect effect 2_ps* a2,b2 -.502 .175 -.840 -.157 Indirect effect 3_ps* a1,d21,b2 .338 .108 .153 .570
Total indirect effect_ps* .087 .219 -.327 .534
Table 4: output summary of mediation analysis in a high monetary compensation setting.
Consequent
M1 (extrinsic) M2 (intrinsic) Y (intention to participate)
Antecedent Coeff. SE p Coeff. SE p Coeff. SE p
X (high vs no) 𝑎1 1.429 .269 p < .001 a2 -.612 .228 p < .001 c' -.277 .159 p = .085 M1 (extrinsic) - - - d21 .348 .085 p < .001 b1 .180 .059 p = .003
M2 (intrinsic) - - - b2 .699 .062 p < .001
constant iM1 3.138 .361 p < .001 iM2 4.675 .339 p < .001 iY .506 .397 p = .205 Control variables Age - - - .013 .004 p = .002 High Income - - - .497 .209 p = .019 -.358 .151 p = .019 Medium Income -.860 .337 p = .012 - - - - R2 = .295 R2 = .240 R2 = .684 F(8,96) = 5.023, p < .001. F(9,95) = 3.337, p = .001 F(10,94) = 20.335, p < .001.
Table 5: summary of direct and indirect effects of mediation analysis for a high monetary compensation setting
Effect Effect_ps* SE p LLCI ULCI
Direct effect c' -.277 -.298 .159 p = .085 -.594 .039
Total effect c -.099 -.107 -.210 p = .639 -.517 .318
BOOT SE BOOT LLCI BOOT ULCI
Indirect effect 1_ps* a1,b1 .277 .117 .082 .534
Indirect effect 2_ps* a2,b2 -.460 .180 -.837 -.117
Indirect effect 3_ps* a1,d21,b2 .374 .110 .167 .602
Total indirect effect_ps* .192 .226 -.243 .640
6. Discussion.
This section addresses the outcomes of hypothesis testing. First, a summary of the results will be provided. Thereafter, results will be comprehended in perspective of the academic literature reviewed in section three. Finally, the limitations of this research layout will be discussed.
6.1. Results summary. 6.1.1. Low monetary setting.
In the case of a low monetary compensation setting, results show that it has, as expected, a direct negative effect on intention to participate. Additionally, extrinsic and intrinsic motivation are related, which leads to partial mediation of the relationship between low monetary compensation and intention to participate. All three indirect effects were significant. The path testing for both extrinsic and intrinsic motivation lead to positive effect on intention to participate. Extrinsic motivation based on a low monetary compensation positively affected willingness to participate. However, intrinsic motivation in a low monetary setting lead to less willingness to participate. This supported the conjecture that intrinsic motivation has a stronger negative effect on intention to participate than the positive effect of extrinsic motivation on intention to participate. Therefore, from the results above, it can be concluded that extrinsic and intrinsic motivation both partially mediate the relationship between low monetary rewards and intention to participate even though there is no negative sequential mediation. This implies that they both are mediators of the relationship but independently of one another.
6.1.2. High monetary setting.
In contrast to expectations, no direct relationship was found between a high monetary incentive scheme and individual’s willingness to participate. Consequently, the main hypothesis that higher compensation leads to more willingness to participate was rejected. However, significant indirect effects stand out and are still relevant for interpretation following Hayes’ argument (2009). The indirect effect of extrinsic motivation based on high compensation had a positive influence on intention to participate, whilst the path testing for intrinsic motivation based on high compensation remained negative for intention to participate contrary to expectations. However, the difference in size between the effect of intrinsic and extrinsic motivation on intention to participate decreased in the high compensation setting in comparison to the low compensation setting. This implies that a higher monetary incentive did have an influence on the both mediators and individual’s intention to participate. Additionally, there was evidence of sequential mediation for intrinsic and extrinsic motivation even though the negative effect was not proven. Therefore, it can be concluded that they both mediate the relationship independently of one another. In conclusion, there is thus no statistically supported
argument that a higher compensation level leads to more willingness to participate whilst intrinsic and extrinsic motivation do mediate that relationship.
6.2. Discussion of the results and perspective on previous literature findings. 6.2.1. Crowding out of intrinsic motivation.
One could say that Titmuss’ intuition was right (1970). As Gneezy and Rustichini (2000), and many others, have proven low monetary incentives have a negative effect on intrinsic motivation. This study is another addition to the literature supporting the theory that low monetary incentives crowd out individual’s intrinsic motivation. In low monetary setting, individuals were more sensitive to intrinsic factors than extrinsic and the introduction of a low monetary reward negatively affected their intrinsic motives which resulted in less willingness to participate. It is however important to note that the effect on extrinsic motivation was positive. Thus, the resulting decrease in intention to participate has to be interpreted as a simple sum, intrinsic motivation’s effect was larger than the one of extrinsic motivation and lead to less willingness to participate.
The implications for a participation income scheme are the following. Intrinsic motivation seems to clearly dominate individuals’ motives of participation. Therefore, it appears as important to frame participation accordingly and ensure individuals understand their participation is meaningful to society and not only a mean to receive monetary compensation. This is a point that Bowles and Polonia-Reyes (2012) underlined as crucial to enhance the effectiveness of monetary incentives. Additionally, it is clear that a PI scheme should not be elaborated on the basis of low monetary rewards or the intended effect will not be reached and participation will be lower than if no compensation at all would have been offered.
6.2.2. Higher incentives, higher performance.
The testing of the effectiveness of higher incentives on intention to participate is more mitigated. There is no evidence that higher monetary incentives lead to more willingness to participate. However, one can argue that there is an underlying effect of increasing the level of compensation. As it was mentioned in section 6.1.2., the difference in effect size between intrinsic and extrinsic motivation on intention to participate decreased in the high compensation level compared to the low. This can be interpreted as a first sign that the effect of a higher compensation is relevant but smaller than expected. Thus, a higher increase in monetary reward would be necessary to lead to significant results. Heyman and Ariely (2004) hypothesized that a higher monetary reward would lead to higher performance, an assumption that is not significantly supported through this study even though there is strong evidence for it. The main contribution of this study would then be the to illustrate the dilemma described by
