These policies have pros and cons. Onthe one hand, they help fight the underrepresentation of some categories whose ability does not differ on average from that ofthe more represented categories (OECD, 2012), and they reduce stereotypes (Beaman et al., 2009). They can also possibly improve the confidence of beneficiaries in the longer run. Onthe other hand, they may generate efficiency losses and resentment if they lead to more able employees being passed over for less able but more protected employees (Holzer and Neumark, 2000) or if no able person can be found (Ahern and Dittmar, 2012). Despite an emerging literature on this topic (Pande, 2003; Besley et al, 2004; Fryer and Loury, 2005; Duflo, 2005; Bertrand et al., 2014), little is known about the causal effectsof reservation policies and their spillovereffects. Laboratory experiments have shown that in a setting where high-performing females shy away from competition (Niederle and Vesterlund, 2007; Datta Gupta et al., 2013), introducing quotas substantially increases females’ competitiveness (Nierdele et al., 2013). The surge in the supply of high-performing individuals to the competitive pool more than outweighs the costs ofthe program. Balafoutas and Sutter (2012) confirm that AffirmativeAction reduces the gender gap without harming male competitors. However, except in the last study showing that post-tournament cooperativeness is not affected and in Leibrandt et al. (2015) who, onthe contrary, found a strong backlash against females when quotas are in use, we know very little
4 The prospects for scaling up climate cooperation nucleating at a small scale, although not necessarily hinging onthe mechanisms of technological spillovers, has received increasing attention in related theoretical work (Ostrom, 2009; Dietz, et al. 2012; Sterner and Damon, 2011; Vasconcelos, et al., 2013; Tavoni, 2013). Network diffusion of behaviors and technology adoption may play an important part in catalyzing cooperation, since adoption by one agent often increases the likelihood that others will become aware of their existence and potential benefits relative to the status quo. Many studies have shown that that mutually reinforcing choices lead to accelerating diffusion of a behavior or to the adoption of a technology once a tipping point has been reached (Granovetter, 1978; Watts, 2002; Weir, 2004). Heal and Kunreuther (2012) focus instead on coordination in games with strategic complementarity, by resorting to the concept of ‘tipping set’, i.e. “a subset of agents who by changing from the inefficient to the efficient equilibrium can induce all others to do the same”. They argue that international climate agreements have these characteristics, and motivate the theory with two often mentioned examples of strategic complementarity: the replacement of leaded gasoline with unleaded gasoline, andthe phasing out chlorofluorocarbons through the Montreal Protocol on Substances that Deplete the Ozone Layer. Both examples show how unilateral action initiated by a subset of actors (in the United States) prompted others to follow suit immediately after 2 . This body of work suggests that unilateral action by a subset of agents might hold promise for promoting widespread cooperation notwithstanding the threat of free riding.
We examine thespillovereffectsof gender quotas onunethicalbehavior: quotas have no effect on men and women’s reporting behavior. However, we find that high performing competitive females are more likely to engage in dishonest reporting in a die-under-cup task with respect to their male counterparts. The mechanism explaining this evidence might be a feeling of redemption associated with being a high performing woman in a male-stereotyped environment. Further research is needed to test the link between tournament-like incentives, the task used, and subsequent unethicalbehavior.
6 as our benchmark treatment testing whether quotas affect effort provision under team incentives in a subsequent real-effort task.
In the Unequal treatment, procedures were the same as in Base except that benefits were split unequally among group members. In particular, the two winners ofthe tournament now each received €0.175 per correctly solved question, while the two losers ofthe tournament received only €0.075 per question that any group member solved correctly. As a result, losers and winners ended up earning different amounts, leading to payoff inequality within the group. This treatment reflects, for example, a situation in which winning the tournament constitutes an internal promotion to a higher rank (e.g. to the role of a team leader) which, in turn, also leads to higher remuneration. Compared to Base, we expect the payoff asymmetry in Unequal to increase the scope for spillovereffectsofthe quota, as it now potentially has direct effectsonthe remuneration in the team task, and hence might increase the perceived unfairness of such policy.
promote women and to reduce this gender gap, have become a popular tool aimed at diminishing this imbalance. Recent experimental studies examined the efficacy of such programs, showing that they indeed have the desired effect of promoting women without harming efficiency (Balafoutas and Sutter, 2012; Villeval, 2012; Niederle et al., 2013; Calsamiglia et al., 2013). Much less is known, however, about potential detrimental effectsofaffirmativeaction policies. The implementation of quotas in tournaments, for example, might create negative spillovereffectson subsequent interactions because they are perceived as unfair. Such situations are relevant, for instance, in firms when members ofthe same team or department compete for a promotion, and afterwards have to collaborate with the very same people again. Discontent about (potentially unfair) promotional decisions that are not solely based on performance might then harm work morale. This, in turn, might negatively affect team performance either in form of reduced effort, or in form of increased adverse behavior. 3 Furthermore, this might also reduce the
The indirect effect is the effect that ego framing exerts on reported performance via ego SC. The corresponding coefficient is the product ofthe two paths making up the indirect effect: (.654)(.580) = .380. As the indirect effect is a multiplication of two direct effects, the assumption of normality usually does not hold significance testing is therefore based on bootstrapping instead of normal-based p-values (MacKinnon, Lockwood, & Williams, 2004). As the current sample size is relatively small, I rely onthe percentile confidence interval rather than the bias-corrected confidence interval, which has been shown to come with inflated type-I-error rates in small samples (Fritz, Taylor, & MacKinnon, 2012; Hayes & Scharkow, 2013). The resulting 95% percentile bootstrap confidence interval (CI) based on 5000 repetitions is [ .012, .891]. As the CI includes zero, I have to conclude that there is no significant mediation effect. However, I acknowledge that this result is onthe edge of significance. In fact, the bias-corrected CI is [.036, 1.018], thus excluding zero and suggesting a significant mediation effect. In light of such ambiguous evidence Fritz et al. (2012) recommend to take into account the direct effects whose product makes up the indirect effect. As illustrated in Figure 3, these effects are both significant, bolstering the
average age of municipal councilors. Our coefficient of interest is negative and significant in all regressions.
Overall, our results show that the introduction of gender quotas decreased the average age of individuals who became politicians in the municipalities affected by the policy: the age ofthe overall political body decreased by more in the treatment group municipalities than in control ones. This result is due to the fact that fewer old male politicians were elected. Given the lack of data on candidates’ lists, we cannot disentangle whether these results are driven by optimizing behaviorof parties or of voters. We now discuss each ofthe two possible mechanisms separately, keeping the other one fixed. As far as parties are concerned, gender quotas impose a constraint on candidates’ choice by party leaders, since lists must now include a lower number of males compared to the unconstrained choice before the reform. Since parties are maximizing the expected number of votes, they must re-optimize the choice of male candidates present onthe list. The significant reduction in the age of male politicians suggests that older men are of a lower expected valence in gaining votes for the party. This is in line with the findings of Lawson et al. (2010) that too old (or too young) candidates have a negative effect on electoral performance. The observed increase in the number of elected young male politicians may also be the result of changes of voting behavior. Even if parties did not alter the age composition ofthe male candidates, by their revealed preferences, voters consider younger male politicians better and thus they voted for them. 10
institution that can trigger retaliation behaviorof employees. Several experimental studies have shown that the procedure accountable for a payment distribution affects subjects’ redistribution decisions reflecting fairness perceptions ofthe applied procedures (e.g., Akba¸s, Ariely, and Yuksel 2019; Blount 1995; Bolton, Brandts, and Ockenfels 2005; Cap- pelen et al. 2007, 2013; Konow 2000). When institutions violate procedural-fairness norms, people may refuse to accept unequal outcomes (Grimalda, Kar, and Proto 2016), may engage in cheating (Gill, Prowse, and Vlassopoulos 2013), lying (Banerjee, Gupta, and Villeval 2018), sab- otage (Ambrose, Seabright, and Schminke 2002; Fehr 2018), or theft (Greenberg 1990) in retri- bution. Moreover, procedural unfairness at the workplace can lower workers’ intrinsic motiva- tion (e.g., Breza, Kaur, and Shamdasani 2018; Cohn et al. 2014; Gächter and Thöni 2010; Heinz et al. Forthcoming), work satisfaction (Breza, Kaur, and Shamdasani 2018), and labor supply (Bracha, Gneezy, and Loewenstein 2015). In contrast to the aforementioned studies, we do not focus on antisocial behavior as a retribu- tive justice motive, or on situations in which it is individually beneficial to act unethically. Instead, we focus onthe consequences of a discriminatory payment scheme, characterized by unequal chances of promotion, on antiso- cial behavior towards innocent coworkers. In our scenario, coworkers are not responsible for experiencing unfair treatment. Whereas other studies look at positive spillovereffects from an institution to prosocial behavior (e.g., Engl, Riedl, and Weber 2018; Galbiati, Henry, and Jacquemet 2018), we provide evidence on neg- ative spillovereffects from unfair procedures in one domain onunethical conduct in another domain. These side effects can provide impor- tant insights that should be factored in when evaluating the efficiency of payment institutions.
The number ofthe mother’s siblings should also be a relevant instrument because the amount of resources that is available per child for investments into education depends substantially onthe number of children in the household. This assumes that parents are constrained in borrowing against their children’s future earnings. A significant effect ofthe number of mother’s siblings on her education in the first stage points to such a borrowing constraint ofthe grandparents. Even though there are no schooling fees and very low or no tuition fees at public educational institutions in Germany, investments into children’s education involve forgone earnings for both the parents andthe children. Parents’ time constraints and limited housing space may impose pressure upon the children to make their own living instead of spending more time on educational investments. Table 2 contains a regression of maternal years of education on a set of dummy variables indicating the number of siblings. We estimate two specifications: with and without demographic controls. 7 The table shows a clear negative relationship between maternal education andthe number of siblings. However, mothers without siblings (omitted category) seem to be special having on average less education than mothers with one or two siblings. Black, Devereux and Salvanes (2005) also find this only child particularity, which disappears when they consider the intact family subsample. For this reason we exclude children of mothers without siblings from our sample to allow for a good linear approximation ofthe relationship between the number of
In particular, here we consider theeffects that sub-groups SQ3 and SQ2 may have suffered because of bigger banks’ strategy, which generally cut their customer loans. We view market share ofthe biggest banks as a proxy for their ability to affect the other banks (Goddard et al., 2007), so that we hypothesize – at least at this stage ofthe analysis – a causal direction from larger to smaller banks. Table 6a and 6b report the evidence onthespillovereffectson respectively SQ3 and SQ2 banks in terms on variation of loans. In particular, for SQ3 banks we notice a potential spillover effect, especially when considering the reduction in terms of total assets ofthe whole banking system in each country. Onthe opposite, the performance achieved by SQ4 banks doesn ’t seem to generate specific effects – except in some specifications when considering the reduction of Loans of larger banks (Table 6a). Similarly, even considering the performance of SQ2 banks, we detect no par- ticular effect deriving from SQ4 banks, whilst we may see a common feature – instead of a spillover effect – if considering the performance of SQ3 banks. Against this, for the whole banking system we detect a little spillover effect when considering the reduction in term of total assets, even if miti- gated by the increase of loans. The overall results from these two tables suggest that the hy- pothesized spillovereffects in terms of transferring of market share don’t seem noticeable. Tables 6c and 6d report the evidence onthe deterioration of asset quality for SQ3 and SQ2 banks respectively. To perform this analysis, we consider that a potential deterioration of credit quality which could be ascribed to the reduction of loans from larger banks needs a proper temporal lag to materialize. In particular, here we consider a two-year lag as an adequate compromise between the period that a potential bad loan in average needs to deteriorate andthe length of our dataset.
lar, indigenous people systematically have less favorable educational and labour market outcomes compared to non-indigenous people (Altman, 2000; Dulleck et al., 2016). Affir- mative action is widly used as a policy tool to reduce these patterns of inequality. Such policies give preferential treatment to specific groups of people to compensate for their disadvantaged trajectory, influenced by socioeconomic background, historical discrimina- tion or stereotypes. For instance, in the United States many programs promote minorities in the labour market and higher education institutions. In India, quota systems are in place to favour representation of women and people from lower castes in government and higher education institutions. In Australia, many universities and employers try to achieve participation of indigenous people, at least equivalent to their representation in the Australian population. Despite its popularity, it is still debated in the theoretical and experimental literature whether affirmativeaction increases or decreases the effort exerted by individuals benefitting from such policies (Coate and Loury, 1993).
Originally both minority and majority status were determined by race, whereby African Americans typically represented the minority group and European Americans, federally classified as Caucasians, the majority group. In addition, by the seventies and eighties, this minority status was determined not simply by race, but also by ethnicity, and gender. Typically, these included African Americans, Hispanics, Native Americans, and women. Still, while European American male citizens, who are mostly white, would not qualify under minority status, most members from minority groups, whether non-white or white, were usually eligible. Also eligible were women. That is to say, while many individuals, whether they belonged to a minority group or not, may have had limited chances to enjoy equal educational opportunities due to their rootedness in an economically and socially disadvantaged group, the social justice “affirmativeaction” intended to bestow excluded non-minority groups. By the same token, although the demographic constitution ofthe United States includes many different ethnic groups, including Asian Americans, Italian Americans, and Americans from the Pacific Islands, the ethnic groups legally designated as minority groups included African Americans, American Indians, and Hispanics in addition to women. The logic informing this classificatory system is based on a common denominator members of these various groups share. What these minority groups and women had in common is that they have historically suffered and continue to suffer discrimination due to racism and sexism. The philosophical and ethical framework which enabled the subsequent evolution ofthe concept of “affirmativeaction” was based onthe premise that social discrimination of specific groups had taken place, and that, therefore, a social solution would constitute the best solution to such a social problem. As the “affirmativeaction” model moved through the seventies and eighties, it was understood
There are two possible reasons for these findings. The first reason for why managers and clerical staff may have different fairness perceptions is that they have different values and sources of information (Joy & Witt, 1992). According to Tata (2000), information asymmetries between management and subordinates can influence their respective perceptions of justice. Managers, for example, are likely to focus onthe strategy for administering AA as an important goal of social justice andthe need for management to comply with the provisions ofthe Employment Equity Act 55 of 1998 andthe Promotion of Equality andthe Prevention of Unfair Discrimination Act 4 of 2000. By contrast, subordinates are likely to focus on their knowledge ofthe implications and long-term effectsof AA for them and other co-workers in the organisation. As Tyler (1989:837) points out, ‘people expect an organization to use neutral decision-making procedures enacted by trustworthy authorities so that, over time, all group members will benefit fairly from being members ofthe group’.
With respect to performance in and after the tournament and how it relates to affirmativeaction, our paper offers three main findings. First, when affirmativeaction favors women, it entails no efficiency costs measured by means ofthe total performance of (advantaged and disadvantaged) group members in the tournament or the performance ofthe two winners ofthe tournament. This extends the results of Balafoutas and Sutter (2012) and Niederle et al. (2013) to the case in which the intervention is not implemented exogenously, but emerges within the group as the result of a vote. Second, individual performances in treatment Color follow a pattern that is distinctly different compared to Gender: here, advantaged group members significantly increase their performance when the discriminating intervention is implemented, while disadvantaged group members significantly decrease their performance. We also show that, measuring tournament efficiency in terms ofthe two winners’ performance, the implementation ofaffirmativeaction leads to an efficiency loss in the color treatment with costly voting. These findings indicate that theeffectsofaffirmativeaction policies depend significantly onthe acceptability ofthe factor that determines whether someone is advantaged or disadvantaged through affirmativeaction. 7 Finally, as far as our measure of post-tournament efficiency is concerned, we find that performance in the post-tournament team task does not differ on aggregate depending on whether the policy is implemented, regardless of treatment.
For each pair of mechanisms, we show an example in Appendix A for which one
mechanism outcome Pareto dominates the outcome ofthe other mechanism. A brief discussion about the different results of Proposition 7 and Theorem 1 is in order. Roughly, Theorem 1 obtains by noting that minority reserves does not waste capacity, thus it Pareto improves on majority quotas. The same intuition does not hold in TTC. While applying TTC, although having minority reserves may help minorities by facili- tating cycles that are otherwise impossible (since a school with minority reserves points to minorities and not to majorities), some cycles formed earlier in the procedure may involve majority students. That is, some majority students can be assigned to a school in which he/she has a very low priority. This in turn can make some minority students, who are not in earlier cycles, worse off. Alternatively, TTC with majority quotas prevents majority students from pointing to a school that has no majority quotas, assuring that some majority students are worse off, and might make minority students better off. For a specific example, see Example 5 in Appendix A .
A standard result in contest theory says that the most inefficient (or least able) agents might not actively participate in the competition (Stein 2002). And indeed, ‘minority rep- resentation’ is an important concern in real competitions. For instance, in California the Disabled Veteran Business Enterprise and Small Business Certification Programs establish explicit target market shares for these disadvantaged groups. Similarly, the European Union has target shares for female representation on firms’ boards and some universities in the U.K. have widening participation programs aiming at broadening the range of students who attend university so that they are representative ofthe home population. The challenge is then to design affirmativeaction policies that can reconcile the conflicting aims of reaching both (i) a sufficient level of minority representation and (ii) a sufficient level of competition. Avoiding trade-offs between these objectives is important because it influences the political support for andthe prevalence ofaffirmativeaction policies. Ayres and Cramton (1996), for example, report that various California ballot initiatives tried to end state-sponsored af-
The anti-discriminatory legal provisions in India are patchy. The Scheduled Castes andthe Scheduled Tribes (Prevention of Atrocities) Act (1989), a modified version ofthe first anti- untouchability Act, 9 recently amended in December 2015 to include new categories of actions to
be treated as offences, including barring entry into an educational institution on account of caste, mainly targets caste-based hate crimes. Offences under this Act are treated as a criminal liability, which means that the burden of proof required is much larger than that for a civil liability, as, indeed, is the punishment, if guilt can be established. However, the conviction rate under this Act is low (the three years from 2012 to 2015 have seen a conviction rate of 30 per cent). This low rate, in addition to reflecting possibly inherent pro-elite and pro-UC biases in the system, reflects a critical difficulty in the use of this law: the aggrieved party has to be able to prove that the crime was committed because of their caste status and not because of any other motive. This is often impossible unless the offence was accompanied by open slurs that establish caste as the key reason for the perpetration ofthe crime. Also, conviction is only possible if a complaint has been filed under this act. Descriptive accounts suggest that even the filing of a complaint—the
Even after controlling for firm fixed effects, year fixed effects, region-specific time effects, and industry-specific time effects, there may still remain differences across firms in factors such as management practices that vary over time and that influence the evolution of minority and female representation at the firm, biasing the estimates ofthe effect ofaffirmativeactionon minority and female representation. To reduce this potential source of bias, Equation (1) also includes controls for a set of observable time-varying firm characteristics that are likely to be correlated with unobservable factors such as management practices and that may influence the effect of contractor status onthe share of protected groups at the firm. For example, large firms are more likely to have formalized personnel policies and recruitment programs that may reduce barriers to the hiring of women and minorities, so one might expect larger firms to have better affirmativeaction track records. 14 As well, one might expect contractor status to be positively correlated with firm size. In this case, a positive revealed relationship between contractor status and growth in female and minority employment shares might be spurious, picking up the correlation between protected group share and firm size. Equation (1) therefore includes controls for firm size and whether the firm is a multiestablishment organization. It also controls for the proportion of white-collar nonclerical employees at the firm, since firms with occupational structures that draw more heavily from the white-collar nonclerical workforce may exhibit smaller growth in female and minority representation because women and minorities are underrepresented in the high-skill labor markets from which these firms hire. 15
Much ofthe popular debate surrounding affirmativeaction in higher education focuses on how it affects the allocation of students to universities, taking the achievement of high school graduates as fixed. However, the disparities in educational preparation which drive racial differences in enrollment at selective colleges arise early in the education pro- cess and are formed well before college admissions come into play. As affirmativeaction was originally conceived to mitigate these gaps in racial achievement, it is natural to ask whether and how the removal of racial preferences affects these gaps. In addition, a num- ber of scholars have pointed out that since policies such as Prop 209 give colleges and universities an incentive to place a greater weight on non-academic factors in determin- ing admissions, they could lower student quality by weakening all students’ incentives to invest in their academic qualifications prior to college entry.
product, technology or idea to who is best at execution ofthe same ( Guillen and García-Canal, 2012; Radjou et al., 2012 ).
In summing up and concluding, traditional theoretical lenses tend to explain established ﬁrm competitiveness through a focus on either position or possession. This raises a challenge when attempting to strategize in situations where the ﬁrm has neither a particularly advantageous position nor a clear resource advantage. Theaction-based perspec- tive that we advance is more dynamic and complements the logics of position and possession, thus broadening our understanding of ﬁrm competitiveness. The argument allows scholars to better understand interﬁrm rivalry and dynamics in fast-changing environments. Particularly in such contexts, it complements as well as improves on previous explana- tions and established theoretical frameworks to help explain the emergence, vitality and expansion of many seemingly disadvantaged ﬁrms in recent years. Of course, an impor- tant question, and major challenge, is whether and how agile ﬁrms are able to maintain their agility as and when they become more established. Opportunities abound for research to extend and test the ideas presented here and we hope that this paper instigates scholars to join this conver- sation and do so.