Self-set salaries and creativity

Brück, Christian; Knauer, Thorsten; Meier, Harald; Schwering, Anja Article — Published Version Self-set salaries and creativity Journal of Business Economics Provided in Cooperation with: Springer Nature Suggested Citation: Brück, Christian; Knauer, Thorsten; Meier, Harald; Schwering, Anja (2020) : Self-set salaries and creativity, Journal of Business Economics, ISSN 1861-8928, Springer, Berlin, Heidelberg, Vol. 91, Iss. 1, pp. 91-121, https://doi.org/10.1007/s11573-020-00985-z This Version is available at: https://hdl.handle.net/10419/288632 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Vol.:(0123456789) Journal of Business Economics (2021) 91:91–121 https://doi.org/10.1007/s11573-020-00985-z 1 3 ORIGINAL PAPER Self‑set salaries andcreativity ChristianBrück1 · ThorstenKnauer1 · HaraldMeier1· AnjaSchwering2 Published online: 6 May 2020 © The Author(s) 2020 Abstract Organizations frequently try to incentivize employees to develop highly creative solutions. In this study, we examine self-set salaries as a specific type of incentive design. We investigate whether self-set salaries affect employees’ motivation and overall (creative) performance. Moreover, because self-set salaries potentially risk opportunistic employee behavior, we consider the effect of the observability of peer performance on employees’ level of self-set salaries. Using a laboratory experiment, we hold the average employee compensation constant and demonstrate that, in comparison with fixed-pay contracts, self-set salaries increase the quantitative performance in creative tasks without affecting the average creativity. However, we do not find significant differences between the amount of individuals’ self-set salaries with observability of peer performance and the amount for individuals without the chance to observe peer performance. Our findings are important for firms that rely on the development of creative ideas but are unsure about the effects of the introduction of self-set salaries. Keywords Creativity· Incentives· Self-set salaries· Employee participation· Autonomy JEL Classification C91· M40· M41 Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s1157 3-020-00985 -z) contains supplementary material, which is available to authorized users. * Christian Brück [email protected] Thorsten Knauer [email protected] Harald Meier [email protected] Anja Schwering anja.schw[email protected] 1 Ruhr University Bochum, Universitaetsstr. 150, 44801Bochum, Germany 2 University ofPotsdam, August-Bebel-Str. 89, 14482Potsdam, Germany 92 C.Brück et al. 1 3 1 Introduction This study examines the effects of self-set salaries as a means of stimulating employees’ motivation to develop creative ideas. More specifically, we investigate whether self-set salaries affect creative performance in terms of quantity and average creativity levels compared with fixed compensation. We compare the effects of self-set salaries and fixed salaries because the economic prediction for the two incentive systems is the same, since compensation is not linked to performance. Moreover, because self-set salaries risk opportunistic employee behavior, we study the effect of the observability of peer performance on employees’ salary requests. All innovation, for example, new product development or new services, begins with a person or a team having a creative idea as a starting point (Amabile etal. 1996). Thus, organizations frequently demand that their employees develop highly creative solutions (Adler and Chen 2011; Fallon and Senn 2006; Speckbacher 2017). At first sight, offering incentives to stimulate and motivate creativity appears to be a feasible way to increase employees’ creativity (Byron and Khazanchi 2012). However, it is still a challenge for the management and management accountants to provide adequate incentive or reward systems to promote creativity successfully (Kaplan and Norton 1996). A growing stream of literature in the field of management accounting investigates different types of performance-contingent incentive systems and their effect on creative performance (e.g., Chen etal. 2012; Kachelmeier and Williamson 2010; Kachelmeier et al. 2008, 2019). Unlike the traditional understanding of performance-contingent incentives that refer to objective measures of productivity, such as the quantity produced (Bonner and Sprinkle 2002; Bonner et al. 2000), these studies investigate the “softer creativity dimension of productivity” (Kachelmeier et al. 2008, 342), meaning that performance-based incentive schemes are contingent on explicit measures of quantity, creativity, or both. These studies show that incentives that reward quantity lead to an increased number of creative ideas but may decrease the average creativity of these ideas. Moreover, incentives that reward creativity or both creativity and quantity lead to highly creative solutions but do not necessarily increase the number of produced ideas. These findings are in line with the reasoning that creativity is not affected simply by exerting more effort (Amabile 1996). Instead, creativity requires a multistage process that ranges from the initial preparation to the detached incubation to the eventual creative gains (Armbruster 1989; Csikszentmihalyi and Sawyer 2014). Incentives concerning quantity may therefore encourage employees to focus on easy solutions instead of exploring new approaches (Amabile 1996; Grabner 2014). However, even when prior research provides evidence for the success of performance-contingent rewards, setting up a creativity-contingent or quantity-contingent reward system in an organization is challenging concerning the objective measurement of both creativity and quantity (Kachelmeier etal. 2008). Based on these considerations, performance-based incentive systems might be rather uncommon in organizational departments that require creativity (Lambert etal. 1993; Sprinkle 2008). 93 1 3 Self-set salaries andcreativity In this study, we investigate an alternative compensation system that has gained in importance in creative firms––self-set salaries. We define self-set salaries as compensation contracts in which employees have the opportunity to determine (parts of) their own compensation. Self-set salaries are used in self-managed organizations that delegate decision rights to their employees (Laloux 2014). For example, the Brazilian manufacturer Semco and the U.S. manufacturer Skyline have delegated at least part of the salary determination process to their employees. In general, selfmanagement is expected to increase employee performance through increased motivation (Harrison and Freeman 2004). Consequently, prior research finds that delegating the salary determination process to employees is positively associated with employee performance (e.g. Charness etal. 2012, 2015; Mellizo etal. 2014). We contribute to prior research and investigate whether self-set salaries also affect creative performance. Self-set salaries may be especially suitable for successfully incentivizing creative performance because, compared with performance-con- tingent rewards, self-set salaries do not explicitly emphasize the quantity dimension of creative performance and may not crowd out creativity. Further, while measuring creativity for remuneration purposes is necessary under performance-contingent rewards, this challenge becomes obsolete when deploying self-set salaries. Finally, compared with fixed payments, compensation is not linked to performance either under self-set salaries. While the economic prediction for both incentive systems is therefore the same, self-set salaries are linked to higher levels of autonomy, competence, and relatedness such that creative performance may increase for psychological reasons. To assess the suitability of self-set salaries for enhancing creativity, we investigate the effects of self-set salaries on creative performance in comparison with fixed payments. We argue that participation in the salary determination process satisfies basic psychological needs and thus increases intrinsic motivation compared with fixed payments (Ryan etal. 1996). Regarding creative performance, employees may direct their motivation toward two different dimensions of effort: quantity of ideas and creativity of ideas. We expect employees to focus on the quantity dimension because quantity is salient and therefore easily measurable and every produced idea induces feelings of completion and competence, which reinforce intrinsic motivation (Ryan etal. 1996). Consequently, we predict that self-set salaries lead to a higher quantity of produced ideas than fixed payments. Regarding the effect of self-set salaries on the creativity of ideas, on the one hand, a focus on quantity may negatively affect creativity if employees focus purely on the more objective dimension of creative performance (Amabile 1996). Nevertheless, self-set salaries also add to an autonomy-supportive work environment that reinforces intrinsic motivation and may foster the development of creative ideas (Amabile and Gitomer 1984; Deci and Ryan 2008; Ryan and Deci 2000). Based on these competing considerations, we posit our second hypothesis in the null form such that self-set salaries have no effect on the creativity of produced ideas compared with fixed payments. For our third hypothesis, we consider that self-set salaries naturally contain the risk that employees set unreasonably high salaries. Specifically, we investigate how information about their peers’ performance affects employees’ salary request. 94 C.Brück et al. 1 3 In fact, most employees do not work in isolation and often observe and compare their results with their peers’ results (Falk and Ichino 2006; Mas and Moretti 2009).1 These observations may affect the evaluation of their own creative performance and their salary requests. Consequently, we investigate salary requests based on whether employees have observed their peers’ performance, that is, the outcome of their work, before submitting a salary request.2 We argue that employees strive for fair input–outcome relations (Adams 1965). Thus, without any information about their peers’ performance, employees are unlikely to request very high salaries because they feel like behaving opportunistically. However, when peer performance is observable, employees are almost always able to consider their own performance as superior. For example, “selective” top performers who produce either many ideas or a few very creative ones can convince themselves that the performance dimension on which they performed well (quantity or creativity) is the more important one. Even low performers are able to consider themselves as smart workers who deserve a high salary because they understood the economic incentives of the compensation system. Therefore, we argue that employees who are able to observe their peers’ performance can justify high salary requests and do not experience negative feelings when claiming a high salary. Consequently, we predict that salary requests are higher when peer performance is observable than when it is not. To test our predictions, we conduct a laboratory experiment in which participants perform the “rebus puzzle” task that isapplied in prior studies on incentives and creativity (Kachelmeier and Williamson 2010; Kachelmeier etal. 2008, 2019).3 Specifically, we ask the participants to produce creative rebus puzzles that use images or signs to represent common phrases or wordplays. Independent raters then judge the creativity of the puzzles created. 1 A large stream of accounting literature investigates the provision of relative performance information (RPI) (for an overview, see Schnieder 2018). While our construct “observability of peer performance” is similar to the provision of (cardinal) RPI, we decisively refrain from labeling our construct “RPI.” Accounting research on RPI often focusses on social comparison, impression management, and consequently employee performance. However, firms commonly avoid high transparency or publicity when it comes to the comparison of creative performance and instead try to foster competition with outside organizations (Amabile 1996). This is because public RPI leads to a competitive atmosphere, which causes employees to focus more on the competition than on creating innovative ideas (De Dreu etal. 2008). Thus, by referring to “observability of peer performance,” we stress the feedback function of peer performance information that we want to investigate in this study. 2 Employees may generally be able to observe both the effort and the performance of their peers. However, since creative performance is considered to be less effort driven (Amabile 1996) and effort is thus less meaningful for the determination of a salary, we only investigate the observability of peer performance. 3 While it would also be possible to measure creativity through psychometric tests, these tests pose several problems. In particular, it is questionable whether individuals who answer predetermined questions correctly really exhibit creative behavior or whether they think creatively (El-Murad and West 2004; Weisberg 1993). Hence, we use a modified “consensual assessment technique” (Amabile 1982) that calculates the average value of a person’s level of creativity assessed by several judges. Thus, we can determine both the participants’ quantity and average perceived creativity of puzzles. 95 1 3 Self-set salaries andcreativity Concerning the manipulations, we deploy a nested design. First, we manipulate the payment scheme between subjects at two levels: self-set salaries and fixed payment. Nested within the self-set salary condition, we manipulate whether participants observe their peers’ performance before making their salary requests. To hold the economic incentives constant, we determine the fixed payment based on the average self-set salary on the first day of the experiment. In line with our prediction, we find that self-set salaries increase the quantity of produced puzzles relative to fixed payments. Further, we do not reject the null hypothesis that self-set salaries have no effect on the creativity of produced ideas compared with fixed payments. In addition, participants request higher salaries when they observe their peers’ performance before the request. However, the difference in salary requests is insignificant. We perform several additional analyses that shed more light on the relationship between self-set salaries, observability of peer performance, and creativity. Specifically, the analyses suggest that individuals’ satisfaction with their own performance (after the comparison) has a significant impact on their salary requests. Our study contributes to both management accounting theory and management accounting practice. From a theory perspective, we add to the literature stream that investigates incentives and creativity (e.g., Brüggen et al. 2018; Chen et al. 2012; Kachelmeier and Williamson 2010; Kachelmeier etal. 2008). We show that self-set salaries are an instrument that increases the quantity without affecting the average creativity. Further, our study contributes to research on self-set salaries (e.g., Charness etal. 2012, 2015; Faillo and Piovanelli 2017; Jeworrek and Mertins 2019). While prior research focuses on non-creative tasks, we provide evidence that self-set salaries increase intrinsic motivation and consequently performance in a creative task setting. With regard to practice, our results are particularly relevant for firms of which the competitive advantage builds on the development of creative ideas. Specifically, our study suggests that firms can increase the quantity of ideas without affecting the creativity of ideas by using self-set salaries. Further, since the subjective performance evaluation of creative work is frequently subject to evaluator biases, which are complex to handle or require calibration committees (Demeré etal. 2019), firms can successfully avoid the problematic measurement of creativity and its link to pay by using self-set salaries. However, along with self-set salaries, additional mechanisms should be implemented that prevent inappropriate salary requests, because our findings show a high number of requests for the maximum salary. The remainder of the study proceeds as follows. Section2 provides the theoretical background and discusses the related literature. In Sect.3, we develop our hypotheses. Section 4 describes the experimental design. Section 5 presents our results, while Sect.6 concludes. 2 Background andrelated literature 2.1 Creativity andincentives Creativity refers to the generation of new ideas, solutions, or products in any domain that are not only novel but also appropriate (e.g., Amabile 1996; Byron etal. 2010; 96 C.Brück et al. 1 3 Shalley and Perry-Smith 2001). The term “appropriate” means that the creation is useful in that it offers value in a practical, esthetic, or intellectual sense (Byron and Khazanchi 2012). Innovation, defined as the successful implementation of creative ideas within an organization, is therefore a consequence of creativity by individuals or teams (Amabile etal. 1996). Motivating creative performance appears to be substantially different from motivating employees to work on routine, structured, or repetitive tasks. The development of creative ideas requires skills such as curiosity, cognitive flexibility, or persistence in the face of barriers (Amabile etal. 1996; Shalley etal. 2004; Zhou and Shalley 2003), which creativity research considers to be unaffected by increases in effort from extrinsic incentives (Amabile 1996; Condry 1977). Instead, several aspects of the organization’s work environment (e.g., flexible working hours, a flying desk policy, and a home office) are considered to affect creative performance (e.g., Amabile 1988; Woodman etal. 1993). The positive effect of these aspects on creative performance derives from enhanced intrinsic motivation because creativity is more cognitively demanding and of less certain value than performing structured tasks repeatedly (Amabile 1996; Eysenck 1995). In fact, individuals are expected to be most creative when they are primarily intrinsically motivated; that is, they experience interest, enjoyment, or satisfaction from the work itself (Amabile 1983, 1988, 1993). Given the link between intrinsic motivation and creativity, prior literature concludes that individuals (or teams) who experience autonomy in terms of freedom or discretion during their work exhibit higher levels of creativity (e.g., Amabile and Gitomer 1984; Bailyn 1985; Paolillo and Brown 1978). If, however, extrinsic incentives lead individuals to feel externally controlled in their work, their intrinsic motivation and thus creative performance are expected to decrease (Amabile etal. 1990; Joussemet and Koestner 1999). Hence, using performance-based incentives to promote creativity may not be appropriate. In contrast to this expectation, a review of the extant studies on creativity and rewards by Byron and Khazanchi (2012) shows that creativity-contingent rewards are able to increase creative performance. However, the authors find that, when rewards are quantity contingent, incentives have a slightly negative effect on creative performance. Further, an emerging stream of management accounting literature examines the relationship between creativity and incentives. For example, Grabner (2014) finds that, in creativity-dependent firms, performance-based pay and subjective performance evaluation are used complementarily. Hence, firms may find performancebased pay to be effective in evoking creativity. Moreover, in an experimental study, Kachelmeier etal. (2008) find that quantity incentives increase the number of ideas and that creativity-based pay increases the average creativity of ideas. However, when both quantity and creativity are incentivized by a creativity-weighted measure, participants perform worse with regard to the weighted measure than participants who are incentivized based on the number of creative ideas. The authors argue that participants focus on the creative dimension of the weighted measure and do not produce enough “mediocre” ideas. In a follow-up study, Kachelmeier and Williamson (2010) find that the possibility to self-select into either a contract that rewards quantity only or a contract that rewards both creativity and quantity strongly affects initial creativity. However, this effect is not sustainable when all the produced ideas 97 1 3 Self-set salaries andcreativity are considered. Kachelmeier etal. (2019) examine the creative process and incentives in this process. The authors argue that the creative process consists of multiple stages––preparation, incubation, and illumination––and that incentives may be especially useful in the preparation stage. In line with this expectation, the results show that incentivizing quantity compared with fixed compensation in the first phase of an experiment leads to more high-creativity solutions in the second phase of an experiment that takes place after an incubation period. Thus, quantity-based measures may increase not only quantity but also creativity after an incubation phase. Finally, Chen etal. (2012) find that a group tournament based on creativity ratings increases group creativity compared with a group piece rate that is also based on creativity ratings. Hence, group tournaments based on creative performance may be a suitable mechanism to increase creative performance in groups. In sum, prior research shows that incentives affect performance with regard to both quantity and creativity. However, while quantity-based incentives are often easier to implement, they potentially hurt the creativity of ideas. In contrast, incentivizing creativity may lead to highly creative performance, but it is challenging to measure the creativity of ideas objectively. Against this background, we investigate an alternative incentive system––self-set salaries. 2.2 Self‑set salaries In general, prior research provides evidence for a positive association between employees’ salary levels and their willingness to perform (Charness etal. 2004; Fehr etal. 1993). However, instead of just raising the salary level to evoke high effort and performance, alternative mechanisms may be suitable for increasing employee motivation. For example, organizations may delegate decision rights to their employees, such as the determination of salaries (Laloux 2014). In this regard, the Brazilian manufacturing company Semco has a long tradition of democratic processes that allow their employees to control their working hours, location, and salary (Semler 1989, 2007). Other companies, such as Skyline or Claravision, also involve their employees in the salary determination process (Charness etal. 2012; Tuna 2008). In a broader sense, wage negotiations between firms and workers’ councils and/or trade unions also reflect democratic structures and employee participation (Franke etal. 2016). Thus, democratic structures that range from workers’ councils to completely self-managed organizations are common nowadays. The main reasoning behind high levels of democracy is that employee participation and self-management increase employee effort and performance, because democracy gives employees autonomy and increases their organizational commitment (Harrison and Freeman 2004; Sliwka 2001). Self-determination theory (e.g., Deci and Ryan 1985; Gagné and Deci 2005) describes the cognitive processes that explain how contextual factors can increase or decrease individuals’ intrinsic motivation and hence their subsequent effort and performance. A fundamental prerequisite for enhancing individuals’ intrinsic motivation is to satisfy the three basic psychological needs to experience feelings of autonomy, competence, and relatedness (Ryan etal. 1996). In 98 C.Brück et al. 1 3 this regard, self-set salaries may be an appropriate means to address all three types of psychological needs. When employees are able to determine their own salaries, they experience a high level of autonomy. Moreover, the salary determination process may enhance their feelings of competence, because they are involved in an essential part of organizational management. Finally, employee participation typically creates a sense of responsibility and relatedness to organizational outcomes. Thus, self-set salaries may be an instrument that increases employees’ intrinsic motivation. Self-set salaries can also circumvent potential crowding-out effects of extrinsic rewards. The crowding-out hypothesis states that extrinsic incentives, such as monetary rewards, can undermine individuals’ intrinsic motivation (Deci and Ryan 1985; Frey and Jegen 2001; Gneezy and Rustichini 2000). Consequently, when extrinsic rewards are not designed appropriately, effort and performance may be even lower under extrinsic rewards than when no rewards are used. However, because self-set salaries do not link performance to pay and address feelings of autonomy, competence, and relatedness instead, intrinsic motivation is likely to be enhanced by this type of compensation system. Prior literature investigates self-set salaries primarily in experimental economics settings. These studies differ with regard to the degree of decision authority delegated to the employees in the salary determination process. Some studies let participants choose from a range of compensation options. For example, Mellizo etal. (2014) show that, when participants are allowed to vote on a set of different compensation schemes, they exhibit higher levels of effort. In contrast, Franke etal. (2016) do not find a positive effect of increasing participation in such a salary determination process. Further studies investigate the effect of entirely self-set salaries. Charness etal. (2012) investigate, in a gift exchange setting, how a firm’s decision to delegate the salary determination process to the employees affects their effort. They find that the decision to delegate significantly increases the effort levels and that this result is driven by participants’ feelings of responsibility rather than positive reciprocity toward the firm that decided to delegate. Further, Charness etal. (2015) show that the positive effect of delegation holds when there is more than one employee and when a real-effort design instead of a chosen-effort design is deployed. Faillo and Piovanelli (2017) provide evidence that the delegation of the salary decision indeed increases intrinsic motivation. In line with these findings, Jeworrek and Mertins (2019) show that delegating the salary choice increases performance significantly in a natural experiment. The authors show that the positive effect is primarily driven by the transfer of responsibility. In conclusion, prior literature shows that self-set salaries increase effort and performance in routine tasks. We contribute to this line of research by investigating the effect of self-set salaries in the context of creativity. Moreover, because employees may also exploit the possibilities of participation by requesting unreasonably high salaries (Harrison and Freeman 2004), we investigate how salary requests may be affected by the observability of peer performance. 105 1 3 Self-set salaries andcreativity the condition with observability of peer performance then learned that they had two minutes to look at the other participants’ puzzles before making their payment request. The experimenter called the participants in pairs, who were then able to look at the puzzles, starting at different sides of the desk and not being allowed to talk to each other. After all participants had assessed the puzzles, they were asked to make their payment requests using the invoice form. In contrast, the participants in the self-set salary condition without observability of peer performance filled out the invoice form directly after the designing phase ended, so they had no chance to learn about their peers’ creative performance. In both conditions, the salary requests were kept confidential so that the participants did not receive any information about the other participants’ salaries. Eventually, both the participants in the self-set salary condition without observability of peer performance and the participants in the fixed-payment condition were allowed to look at the other participants’ puzzles after all experimental decisions had been made. Thereby, the same procedures were applied as described above. Thus, we ensured that all participants had the same state of knowledge about the performance of all participants in the respective session before filling out the PEQ. The PEQ asked questions related to the experimental procedures, the participants’ personality, and general demographics. Finally, participants were instructed about their total experimental compensation and the payout procedures. Figure1 provides an overview of the experimental procedures. Fig. 1 Experimental procedures 106 C.Brück et al. 1 3 4.4 Participants We randomly assigned 89 participants to the three conditions (30 participants in both self-set salary conditions with/without observability of peer performance and 29 participants in the fixed-payment condition). We recruited undergraduate and graduate students from a large Western European university to participate in 70-min experimental sessions, from which we obtained 83 usable observations.10 In the final sample, 59.8 percent of the participants majored in a business-related subject, 62.2 (37.8) percent were undergraduate (graduate) students, 59.9 percent of participants were female, and the participants’ average age was 24.4years. 4.5 Creativity ratings For our hypothesis tests, we use three different dependent variables: the quantity of designed puzzles, the average creativity of designed puzzles, and the requested payment. While the quantity of designed puzzles and the requested payment can be determined objectively, we assessed the creativity of the puzzles through independent evaluations. Specifically, we followed prior experimental studies (e.g., Brüggen etal. 2018; Kachelmeier and Williamson 2010) and invited twelve undergraduate and graduate students from all fields of study to rate the creativity of all 850 rebus puzzles created by the experimental participants. The raters were paid 50 euros for their time. We randomized the order of the 850 rebus puzzles such that the raters were not able to identify any patterns or abilities related to a specific participant. First, all the raters had to read the instructions on the experimental task given to the experimental participants. However, they were not informed about the different compensation schemes. Afterward, the raters evaluated the creativity of every puzzle by entering a score between 1 (lowest creativity) and 10 (highest creativity). The raters were advised to assess the creativity of about 20 puzzles first before starting to enter scores. Thereby, they would gain a better idea of the creativity of the puzzles. The Cronbach’s alpha for the rating of all ideas is 0.86. Moreover, correlations between each rater and the average of the other eleven raters are significantly positive (ranging from 0.45 to 0.71). We therefore conclude that the creativity ratings are reliable and consistent across the raters. To determine the average creativity of the rebus puzzles for every participant, we first average the ratings for every puzzle across raters. Thus, we obtain an average rating per puzzle. Next, we calculate average creativity for every participant as the sum of the average ratings of submitted puzzles divided by the number of submitted puzzles. 10 As outlined in Sect.5.1, six participants failed the manipulation checks. 107 1 3 Self-set salaries andcreativity 5 Results 5.1 Experimental design validation To assess the validity of our experimental design, we test for successful randomization of participants across experimental conditions. Regarding the participants’ demographics, we find that there are no significant differences across conditions regarding age, gender, pursued degree, or field of study (all p values > 0.42). Moreover, Kachelmeier et al. (2008) find in their study that familiarity with rebus puzzles is positively correlated with the average creativity ratings. Therefore, we asked in the PEQ the extent to which the participants agree (7-point Likert scale) with the statement that they had some experience with rebus puzzles prior to this experiment. However, we find that the familiarity with rebus puzzles influences creativity in terms of neither quantity (F = 1.56, p = 0.22) nor average creativity (F = 1.51, p = 0.22). Further, the Kruskal–Wallis equality-of-populations rank test reveals no significant differences in familiarity with rebus puzzles across all three conditions (χ2 = 0.11, p = 0.95). We also analyze differences in personality traits across conditions. More specifically, we test for differences in modesty, which refers to self-crit- ical attitudes. We measure modesty based on the respective questions of the NEO Personality Inventory. Again, the Kruskal–Wallis test does not indicate significant differences across conditions (χ2 = 2.59, p = 0.27). Hence, we conclude that randomization was successful. Further, we test whether our manipulations were successful. In the experiment, we manipulated whether participants were enabled to set their payment amount independently or received a fixed payment. Hence, we asked the participants whether they were allowed to decide on their payment on their own. All participants answered this question correctly. Nested within the self-set salary condition, we manipulated whether participants received peer performance information prior to the payment request. Therefore, we asked the participants in the conditions with a self-set salary whether they determined their payment after they had seen the puzzles of the other participants. Six participants answered this question incorrectly.11 Consequently, we exclude these participants from all analyses. 5.2 Descriptive statistics Table 1 presents descriptive statistics by the experimental cells. Concerning the quantity of designed puzzles, the participants created, on average, 9.59 rebus puzzles. Further, we find that the participants in conditions of self-set salaries with observability of peer performance (10.29) and without observability of peer performance (10.17) created more puzzles than the participants in the fixed-payment condition (8.41), which is in line with H1. Regarding the average creativity, results are 11 Those six participants were all in the observability of peer performance condition and indicated that they did not determine their payment after they had seen the others’ puzzles. 108 C.Brück et al. 1 3 similar across conditions. When combining the two conditions of self-set salaries, the average creativity (5.57) is equal to the fixed-payment condition.12 This result suggests that neither the autonomy-supportive task context improves creativity nor the self-set salaries erode creativity while simultaneously increasing the quantitative output. Finally, Table1 reports the participants’ payment requests. On average, the participants with self-set salaries but without observability of peer performance requested 17.63 euros, which we use as the equivalent payment in the fixed-payment condition by design. The participants with self-set salaries and observability of peer performance requested a slightly higher amount of 17.92 euros. Thus, the participants requested relatively high salaries, which is also reflected in the high number of participants who asked for the maximum amount possible, that is, 20.00 euros (without observability of peer performance: 18 out of 30 participants; with observability of peer performance: 17 out of 30; not tabulated). 5.3 Hypothesis tests Table2 reports the results of our formal hypothesis tests. Hypothesis H1 posits a positive association between self-set salaries and the number of produced ideas in a creative task setting. To test this hypothesis, we pool our treatments of self-set salaries with and without observability of peer performance and compare them with the fixed-payment condition. With respect to the small number of observations, we perform both a t-test and bias-corrected and accelerated bootstrapping (95% confidence interval, 1,000 replications).13 The results from the t-test (t = 2.07, p = 0.042, two-tailed) and the bootstrapping (z = 2.17, p = 0.030, two-tailed) support hypothesis H1, indicating that self-set salaries increase the number of produced puzzles. 13 When testing whether the requirements of the t-test are met, a skewness test for normality reveals that participants’ quantitative output is significantly skewed (p = 0.002). Performing a Shapiro–Wilk W test for normality confirms this pattern and indicates non-normal data (W = 0.949, p = 0.002). Therefore, we use logarithmic values to perform our analyses, which solves the issue of skewed results. Both the skewness test (plog = 0.517) and the Shapiro–Wilk W test for normality (Wlog = 0.992, p = 0.881) indicate that our data follow a normal distribution after the log-transformation. Table 1 Descriptive statistics The table presents the mean values (standard deviation) for each dimension and condition Self-set salaries Fixed payment (n = 29) Observability of peer performance (n = 24) No observability of peer performance (n = 30) Total (n = 83) Quantity 8.41 (3.17) 10.29 (3.54) 10.17 (4.24) 9.59 (3.75) Average creativity 5.57 (0.98) 5.46 (0.97) 5.66 (1.14) 5.57 (1.03) Payment request 17.63 (0.00) 17.92 (3.06) 17.63 (3.80) 17.71 (2.79) 12 Likewise, participants’ highest-rated rebus puzzles are in a similar range, with means of 7.14, 7.02, and 7.11 in the fixed-payment condition, the self-set salary condition with observability of peer performance, and the self-set salary condition without observability of peer performance, respectively. 109 1 3 Self-set salaries andcreativity Table 2 Hypothesis tests Panels A to C of this table present the mean values (standard deviation) per condition or the test statistics (two-tailed p-values), respectively a The t-test results are based on logarithmic values b We bootstrap our results with bias-corrected and accelerated (BCa) confidence intervals (1000 replications) Panel A: Hypothesis 1 Variable Self-set salaries (n = 54) Fixed payment (n = 29) t testaBCa bootstrapb Quantity 10.22 (3.91) 8.41 (3.17) 2.07 (0.042) 2.17 (0.030) Panel B: Hypothesis 2 Variable Self-set salaries (n = 54) Fixed payment (n = 29) t test BCa bootstrapb Average creativity 5.57 (1.06) 5.57 (0.98) 0.01 (0.992) 0.01 (0.992) Panel C: Hypothesis 3 Variable Self-set salaries without observability of peer performance (n = 30) Self-set salaries with observability of peer performance (n = 24) t test BCa bootstrapb Payment request 17.63 (3.80) 17.92 (3.06) 0.30 (0.768) 0.31 (0.759) 110 C.Brück et al. 1 3 Regarding hypothesis H2, we use the same research design to investigate whether self-set salaries affect the average creativity differently from fixed payments. As indicated by the descriptive results, we do not find significant differences in the average creativity when using either the t-test (t = 0.01, p = 0.992, two-tailed) or bootstrapping (z = 0.01, p = 0.992, two-tailed). This finding indicates that self-set salaries neither harm nor benefit the creative dimension of creative performance. Thus, in contrast to the results of Kachelmeier etal. (2008), who show that quantity-contin- gent incentives increase the quantity but reduce the creativity, we find that, while the quantity increases, the creativity remains unchanged when self-salaries instead of fixed payments are used. Regarding hypothesis H3, we expect that the observability of peer performance before the payment request enables individuals to justify their payment requests, which results in higher payment requests. With respect to the descriptive results, the mean payment request of the participants with observability of peer performance is higher than the request of their counterparts without observability of peer performance. However, the difference is insignificant, both for the t-test (t = 0.30, p = 0.768, two-tailed) and for the bootstrapped results (z = 0.31, p = 0.759, twotailed). Therefore, we reject H3. A possible explanation for the lacking support of our H3 could be a ceiling effect. In fact, 31 out of 54 participants (57.4%) requested the maximum amount of 20 euros. The corresponding chi-squared test also determines a uniform distribution of the payment requests between the two self-set salary conditions (χ2 = 7.37, p = 0.497, not tabulated), so our research design may not allow such analyses. 5.4 Additional analyses 5.4.1 Robustness checks To investigate the reliability of our results, we first repeat our analyses using data from all the participants (including the six participants who failed the manipulation checks). Table3 reports the results. We find that the results are inferentially identical to those of our main analyses. Further, we investigate whether the amount of the requested payment interferes with our results. Table4 presents the results of the OLS regression analysis, in which the quantity (Panel A), the average creativity (Panel B), or the creativity-weighted productivity (Panel C) serve as dependent variables. Further, we add the payment request as a control variable. However, the level of the requested payment does not have a significant effect on the respective creativity measures, while self-set salaries have a significant effect on the quantity (p = 0.044, two-tailed) and creativityweighted productivity (p = 0.053, two-tailed). Therefore, our results are supported. 5.4.2 Alternative measures ofcreativity In this section, we report alternative measures of creativity to gain a better understanding of the effects of self-set salaries on creativity. Specifically, Kachelmeier 111 1 3 Self-set salaries andcreativity Table 3 Hypothesis tests (failed manipulation checks included) Panels A to C of this table present the mean values (standard deviation) per condition or the test statistics (two-tailed p-values), respectively a The t-test results are based on logarithmic values b We bootstrap our results with bias-corrected and accelerated (BCa) confidence intervals (1000 replications) Panel A: Hypothesis 1 Variable Self-set salaries (n = 60) Fixed payment (n = 29) t-testaBCa bootstrapb Quantity 10.12 (3.74) 8.41 (3.17) 2.12 (0.037) 2.12 (0.034) Panel B: Hypothesis 2 Variable Self-set salaries (n = 60) Fixed payment (n = 29) t-test BCa bootstrapb Average creativity 5.48 (1.05) 5.57 (0.98) 0.37 (0.716) 0.40 (0.692) Panel C: Hypothesis 3 Variable Self-set salaries without observability of peer performance (n = 30) Self-set salaries with observability of peer performance (n = 30) t-test BCa bootstrapb Payment request 17.63 (3.80) 17.90 (3.09) 0.30 (0.767) 0.29 (0.771) 112 C.Brück et al. 1 3 etal. (2008) also investigate creativity-weighted productivity, which is equivalent to the multiplicative product of quantity and average creativity. Consequently, we investigate whether creativity-weighted productivity differs between conditions with a self-set salary and those with a fixed payment. As shown in Table 5, Panel A, we find that participants with self-set salaries achieve higher levels of productivity (56.13 compared with 46.84, on average).14 The difference is significant, both on a Table 4 Robustness check (effect of the payment level on creativity measures) Panels A to C of this table present the regression results for the dependent variables quantity, average creativity, and creativityweighted productivity and the independent variables self-set salaries (yes/no) and participants’ payment requests. The values for quantity and creativity-weighted productivity are logarithmic Independent variables Coef. Std. Err. t p-value Panel A: Quantity (dependent variable)  Constant 1.71 0.28 6.17 < 0.001 Self-set salaries 0.18 0.09 2.05 0.044  Payment request 0.02 0.02 1.35 0.180 R20.072  Adjusted R20.048 F (2, 80) 3.08 p-value 0.051 N 83 Panel B: Average creativity (dependent variable) Constant 5.17 0.75 6.86 < 0.001 Self-set salaries − 0.001 0.24 < − 0.001 0.998 Payment request 0.02 0.04 0.55 0.581 R20.004  Adjusted R2− 0.021 F (3, 50) 0.15 p-value 0.858 N 83 Panel C: Creativity-weighted productivity (dependent variable) Constant 3.29 0.30 11.02 < 0.001 Self-set salaries 0.19 0.10 1.96 0.053 Payment request 0.03 0.02 1.62 0.110 R20.076 Adjusted R20.053 F (2, 80) 3.31 p-value 0.042 N 83 14 Similar to participants’ quantity, the results are skewed when testing for normality (p < 0.01). Therefore, we perform a log-transformation so that the results do not deviate significantly from a normal distribution (plog = 0.13). 113 1 3 Self-set salaries andcreativity Table 5 Alternative measures of creativity Both panels of this table present the mean values (standard deviation) per condition and the test statistics (two-tailed p-values), respectively a The t-test results are based on logarithmic values. b We bootstrap our results with bias-corrected and accelerated (BCa) confidence intervals (1000 replications) Panel A: Creativity-weighted productivity (quantity × average creativity) Variable Self-set salaries (n = 54) Fixed payment (n = 29) t-testaBCa bootstrapb Creativity-weighted productivity 56.13 (21.56) 46.84 (20.85) 1.98 (0.051) 1.98 (0.048) Panel B: Number of high-creativity ideas Variable Self-set salaries (n = 54) Fixed payment (n = 29) t-testaBCa bootstrapb Number of high-creativity ideas 5.31 (3.09) 4.34 (3.53) 1.30 (0.199) 1.28 (0.202) 114 C.Brück et al. 1 3 parametric basis (t = 1.98, p = 0.051, two-tailed) and when bootstrapping our results (z = 1.98, p = 0.048, two-tailed). This result further stresses that the increase in quantity based on self-set salaries is not negatively affected by a decrease in creativity. Since some firms might be interested in promoting high-creativity solutions, we also investigate the effect of self-set salaries on the number of high-creativity puzzles. We consider a puzzle as highly creative when its mean rating is higher than the median (5.57). As shown in Table5, Panel B, we find that the participants created more high-creativity puzzles in the self-set salary conditions (5.31 puzzles) than in the fixed-payment condition (4.34 puzzles). This pattern is in line with our theory that predicts a general effect of self-set salaries on the quantity of produced ideas. The difference, however, is statistically insignificant (t-test: t = 1.30, p = 0.199, twotailed; bootstrap: z = 1.28, p = 0.202, two-tailed). Yet, this finding could also be due to our experimental design, which may not allow for sufficient variance between conditions. Specifically, participants had only 20 min of time to produce creative ideas. This limits potential differences across conditions because high-creativity puzzles are a fraction of all created puzzles. Therefore, in the real world, where employees work much longer to produce creative ideas, our theory may still hold. 5.4.3 Investigation ofunderlying mechanisms Next, we provide further insights into the underlying process driving our results using the participants’ answers to the PEQ. Table6 displays the mean response to the questions by experiment cell. The post-experimental questions are based on a 7-point Likert scale from 1 (does not apply at all) to 7 (applies fully). Overall, Table6 reveals that the participants in the self-set salary conditions are more satisfied with their performance (satisfaction with own performance) than the participants with fixed pay (z = 1.87, p = 0.062, two-tailed, not tabulated), which is in line with the greater quantity of designed puzzles.15 Further, in developing our hypotheses on the effects of self-set salaries, we argue that employees experience higher levels of autonomy than those receiving fixed payments, which is likely to increase their intrinsic motivation. Therefore, we measure autonomy based on the autonomy subscale of the Basic Psychological Need Satisfaction at Work Scale, developed by Deci etal. (2001). We perform principal-com- ponents factor analysis to compute a composite measure of autonomy.16 The resulting autonomy factor has an eigenvalue of 2.42 and an explained variance of 48.4 percent. We find that the participants in the self-set salary conditions experience higher levels of autonomy in the experiment than the participants in the fixed-pay- ment condition (z = 1.65, p = 0.099, two-tailed, not tabulated). Hence, we conclude that self-set salaries can increase feelings of autonomy. 15 Table 6 also shows participants’ focus on both quality and quantity. While all the participants indicated that they attached more importance to the quality than to the quantity of the rebus puzzles (t = 2.80, p < 0.01, two-tailed), their statements do not vary across conditions—both for quality (χ2 = 2.60, p = 0.273, two-tailed) and for quantity (χ2 = 1.20, p = 0.548, two-tailed). 16 We exclude items 5 and 7 due to low indicator reliability. 121 1 3 Self-set salaries andcreativity Kunda Z (1990) The case for motivated reasoning. Psychol B 108:480–498 Lambert RA, Larcker DF, Weigelt K (1993) The structure of organizational incentives. Admin Sci Q 38:438–461 Laloux F (2014) Reinventing organizations: a guide to creating organizations inspired by the next stage in human consciousness. Nelson Parker, Brussels Liu D, Chen XP, Yao X (2011) From autonomy to creativity: a multilevel investigation of the mediating role of harmonious passion. J Appl Psychol 96:294–309 Mas A, Moretti E (2009) Peers at work. Am Econ Rev 99:112–145 Mazar N, Amir O, Ariely D (2008) The dishonesty of honest people: a theory of self-concept maintenance. J Market Res 45:633–644 McCrae RR, John OP (1992) An introduction to the five-factor model and its applications. J Pers 60:175–215 Mednick SA (1962) The associative basis of the creative process. Psychol Rev 69:220–232 Mellizo P, Carpenter J, Matthews PH (2014) Workplace democracy in the lap. Ind Relat J 45:313–328 Morris S (1983) Omni games: the best brainteasers from Omni magazine. Rinehart & Winston, New York Oldham GR, Cummings A (1996) Employee creativity, personal and contextual factors at work. Acad Manage J 39:607–634 Paolillo JG, Brown WB (1978) How organizational factors affect R&D innovation. Res Manage 21:12–15 Ryan RM, Deci EL (2000) Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. Am Psychol 55:68–78 Ryan RM, Sheldon KM, Kasser T, Deci EL (1996) All goals were not created equal: an organismic perspective on the nature of goals and their regulation. In: Gollwitzer PM, Bargh JA (eds) The psychology of action: linking motivation and cognition to behavior. Guilford, New York, pp 7–26 Semler R (1989) Managing without managers. Harvard Bus Rev 67:76–84 Semler R (2007) Out of this world: doing things the Semco way. Global Bus Org Excell 26:13–21 Shalley CE, Perry-Smith JE (2001) Effects of social-psychological factors on creative performance: the role of informational and controlling expected evaluation and modeling experience. Org Behav Hum Dec 84:1–22 Shalley CE, Zhou J, Oldham GR (2004) The effect of personal and contextual characteristics on creativity: where should we go from here? J Manage 30:933–958 Sliwka D (2001) On the costs and benefits of delegation in organizations. J Inst Theor Econ 157:568–590 Speckbacher G (2017) Creativity research in management accounting: a commentary. J Manage Account Res 29:49–54 Sprinkle GB (2008) Discussion of: measuring and motivating quantity, creativity, or both. J Account Res 46:375–382 Torrance EP (1974) Torrance tests of creative thinking. Personnel Press, Lexington Tuna C (2008) Pay, your own way: firm lets workers pick salary. Wall Street J 7:86 Weisberg R (1993) Creativity, beyond the myth of genius. Freeman, New York Woodman RW, Sawyer JE, Griffin RW (1993) Toward a theory of organizational creativity. Acad Manage Rev 18:293–321 Zhou J (1998) Feedback valence, feedback style, task autonomy, and achievement orientation: Interactive effects on creative performance. J Appl Psychol 83:261–276 Zhou J, Shalley CE (2003) Research on employee creativity: a critical review and directions for future research. In: Martocchio J (ed) Research in personnel and human resource management. Elsevier, Oxford, pp 165–217 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.