The Effect of Intercultural Interactions on Unethical Behaviors

The Effect of Intercultural Interactions on Unethical Behaviors

Name of student: Thanaphat Thongpaibool 11120096 Name of supervisor: Dr. Shaul Shalvi

Name of co-assessor: Dr. Jörg Gross

ABSTRACT

Corruption, as well as other human unethical behaviors, can be observed in every society. As culture is pervasive and entrenched into human society, sizeable empirical evidence has repeatedly tested and reaffirmed the impact of culture and cultural difference on individuals‘ corrupt behaviors. Nevertheless, to my knowledge, none has yet investigated how the interactions between individuals from different cultures may influence individuals‘ corrupt behaviors. On this account, recently, substantial research attention has been paid to investigating the effect of intercultural interactions on economic behaviors, such as cooperation, trust, reciprocity and fairness consideration. In this review, I explore these issues by investigating the pertinent existing literatures in the fields of unethical behaviors and various related economic behaviors, especially on the topic of the effect of intercultural interactions. By synthesizing the explanations and implications from the economic studies, I put forward several hypotheses regarding how the intercultural interactions may influence individuals‘ corrupt behaviors specifically in the sequential dyadic die-rolling game, on the basis of three possible explanatory variables: the country‘s corruption level, the country‘s level of trust, and the cultural value of individualism versus collectivism.

1. Introduction

In modern times, the advent of rapidly growing globalization creates international interdependence between various socio-cultural and economic institutions all over the world. Globalization opens the door for individuals to not only engage and cooperate with fellow compatriots but also encourage frequent international interactions between individuals from different cultural backgrounds—through e.g. migration and international trade and investment. An apparent issue concerning international/intercultural interactions is that individuals from different cultural backgrounds, to a certain extent, hold specific dispositions, such as social norms, beliefs, values, and behavioral patterns. Cultural institutions influence social preferences thereby shaping individuals‘ behaviors (Kim, Choi, Lee, Lee, & Choi, 2013). In this way, misunderstanding between parties of different cultures could occur due to country-specific or culture-specific conducts, which might in turn create unintended conflicts (Hennig-Schmidt, Selten, Walkowitz, Winter, & Dakkak, 2007). Therefore, in the age of globalization, understanding how individuals from different cultural backgrounds behave differently and how the intercultural interactions affect individuals‘ behaviors become more and more important.

In the past decades, ample amount of empirical evidence from cross-cultural experiments using economic games have advanced our understanding on how macro-components, e.g. culture, affect behaviors such as trust, reciprocity, fairness, and cooperation—at the individual level (see Fehr & Camerer, 2002). On top of that, more recent studies have increasingly started to explore not only the individual performance differentials between different cultures, but also economic decisions in the intercultural context (e.g. Fershtman & Gneezy, 2001). Whereas these types of investigations are meaningful in discerning the effect of cultural differences and intercultural interactions on behaviors in economic situations, little attention has been paid to

other domains of behavior, specifically (un-)ethical behaviors. In the context of experimental task or game, corrupt behavior would denote a situation in which individuals do not follow the rules of the task or game, i.e. cheating or lying. To my knowledge, only few existing studies (Gächter & Schulz, 2016; Gelbrich, Stedham, & Gäthke, 2016; Mann, Garcia-Rada, Hornuf, Tafurt, & Ariely, 2016) have probed into investigating the impact of cultural difference on corrupt behavior. More surprisingly, none has yet identified how intercultural interactions may influence unethical behavior. In other words, the unresolved research question, as of now, is: how would intercultural interactions (between heterogeneous groups of individuals from different cultural backgrounds) affect corrupt behavior? To answer this question, here I examine how or whether different cultural backgrounds impact individual‘s corrupt behavior when interacting with one another by reviewing and integrating existing knowledge from pertinent literatures on the role of intercultural interactions on economic behaviors and experimental works on corrupt behavior.

The remainder of this article is organized as follows: The next section provides a review of empirical works in the domain of unethical behavior, specifically corruption, at the individual level and interpersonal interaction, and whether individuals are affected by others when exposed to others‘ corrupt behavior as well as relevant explanatory psychological variables. In the third section, I primarily define the term ‗culture‘ and discuss the impact of culture and cultural difference on corrupt behaviors. In the forth section, I distinguish between cross-cultural and intercultural experiments, explore the topic of the intercultural interaction effect on behaviors, and provide empirical examples in the context of economic decisions. I further review possible determining psychological factors for how intercultural interactions may influence individual economic behaviors in the domain of intergroup cooperation, trust, reciprocity, and fairness. In

the fifth section, I extrapolate on how intercultural interactions may affect individual corrupt behavior in a specific experimental task in relation to possible explanations derived from economic literatures. Finally, the last section concludes the article, considers shortcomings of research relating to culture, as well as suggests research questions for future investigation.

2. Ethical framework: Corrupt behaviors

Unfortunately, corruption is a commonly observed behavior in humans. Although corruption has many different definitions (e.g. Gelbrich et al., 2016), here I circumscribe corruption specifically as a situation in which at least two or more individuals jointly act in collaboration and do not follow or violate the (socially accepted) rules—and that includes dishonest behaviors such as lying, deceiving and cheating. In the context where the outcome of individual acts is independent of others‘, I shall use dishonest behavior interchangeably with corrupt behavior. In an attempt to explain the emergence of corrupt behavior, there are two major schools of thought from classical economics and social psychology. The first follows the canonical economic account which assumes that humans are self-serving and profit-maximizing agents (the so-called homo economicus), and thus, will likely engage in corrupt behavior insofar as the material reward (e.g. money) is concerned (Ostrom, 1998). However, a tremendous body of empirical evidence seems to gainsay this classical assumption and render it insufficient in many circumstances (see Rosenbaum, Billinger, & Stieglitz, 2014). In contrast, an alternative explanation stems from the field of social psychology, suggesting the role of norms in moderating corrupt behaviors (Somanathan & Rubin, 2004). Social norms are collectively shared understandings about which actions and behaviors are socially acceptable or appropriate within a

specific social context and a certain social group (Barr, Lane, & Nosenzo, 2015; Ostrom, 2000). Social norms can be categorized into two types: descriptive and injunctive norms (Gino, Ayal, & Ariely, 2009; Keizer et al. 2008). Descriptive norms prescribe what most people commonly do in a particular setting, whereas injunctive norms specify a particular set of behaviors which most people approve or disapprove of. The degree to which each norm has an effect on individuals‘ immediate behavior is dependent on the social context (Gino et al., 2009). Essentially, this social norms account attests that dishonesty and corruption (e.g. lying) require a psychological cost since doing so violates a social norm accepted by a considerably large number of members within society.

In an experimental setting the scope of corrupt acts is varyingly embraced and a variety of tasks/games have been used to investigate corrupt behaviors. Among the 63 economics and psychology papers pertaining to individual dishonest behavior experiments that Rosenbaum et al. (2014) reviewed, six prominent categories of recurring experimental tasks were identified: the lost or misdirected items (e.g. Franzen & Pointner, 2013; Keizer et al., 2008; West, 2005), undeserved money or windfalls (e.g. Bickman, 1971; Rosenbaum, Billinger, Stieglitz, Djumanov, & Atykhanov, 2012), sender–receiver games (e.g. Robert & Arnab, 2013), ―honor stands‖ (e.g. Bateson, Nettle, & Roberts, 2006; Pruckner & Sausgruber, 2013), self-reported outcomes (e.g. Shalvi, Dana, Handgraaf, & De Dreu, 2011), and theft tasks (e.g., Mazar, Amir, & Ariely, 2008). As such, even among experimental literatures, the range of what is considered as corrupt behavior may be broadly interpreted. For this reason, heterogeneity in research methodologies may contribute to a variety of different conclusions from disparate experiments (see Rosenbaum et al., 2014).

While numerous demographic factors—such as gender, age and study major—provide contradictory results as an explanatory variable, despite the experimental variations, three common salient effects on corrupt behaviors can be robustly identified across the research findings. First, the intrinsic psychological cost of dishonesty has been consistently found to influence the magnitude of dishonest behavior. Shalvi and colleagues (2011), for example, have experimentally shown that the degree of corrupt behavior, specifically lying, is dependent on the extent to which such lies can be self-justifiable—e.g. in the form of counterfactuals. The results also implicitly suggest that the psychological distinction between morally right and wrong is continuous rather than discrete (Lewis et al., 2012). Second, not only does the magnitude of corrupt behavior depend on an internal cost of lying or cheating, an external social cost (such as social appearance) comes into play as well. Specifically, it has been found that a presence of the third party, or even bystanders, mitigates dishonest behavior. In other words, when individuals are being monitored, they tend to act more honestly. For instance, as early as in the Seventies, Steinberg, McDonald and O'Neal (1977) reported that participants stole a luxury shampoo more when they were alone than when there was a bystander. In this case, the reason participants behaving less corruptly when being monitored might be because they want to preserve their positive self-conception of being honest (Fischbacher & Föllmi-Heusi, 2013; Mazar et al., 2008). Lastly, dishonesty level decreases when participants are primed to moral norms prior to the experiments. A field study on the purchase of street newspapers provides a good example, suggesting that a moral reminder based on an internalized social norm ameliorates the level of corrupt behavior measured by the payment amount (Pruckner & Sausgruber, 2013). Accordingly, this seems to suggest that moral decisions may substantially be driven by societal factors, such as

social norms and social preferences, which thereby have an effect on the degree of individuals‘ corrupt behavior (Cappelen, Sørensen, & Tungodden, 2013).

Along the similar line, it has been shown that individual corrupt behavior could potentially be affected by others‘ unethical actions in other situations, even when the outcomes are independent of one another. Observing social norm-violating acts seems to exacerbate the degree of individuals‘ corrupt behaviors. For example, Keizer, Lindenberg, and Steg (2008) conducted six field experiments and concluded that when participants observed others violated a certain contextual norm or rule, which the participants could observe being obviously violated, they tend to violate other norms or rules as well. Specifically, in one of the experiments, littering behavior in a shopping area in Groningen was compared between the two conditions of the anti-graffiti norm. In one condition, despite the obvious disapproval sign for anti-graffiti, the wall was noticeably scribed and painted with graffiti. In contrast, the wall in a control condition was completely clean. Interestingly, 69 percent of participants in the graffiti condition littered compared to 33 percent of participants in the control condition. This seems to suggest that—as a certain norm-violating behavior becomes more prevalent and common—individuals‘ unethical acts can be systematically influenced by others‘ observable behaviors and social norms.

In the intergroup paradigm, the effect of other individuals‘ corrupt behavior seems to be profoundly conditional on social contexts. Through a series of the theft task experiments, Gino, Ayal, and Ariely (2009) tested the effect of exposure to other individuals‘ corrupt behavior based on three psychological determinants: the estimated likelihood of being caught calculated by the cost-benefit analysis, social norms, and saliency of (un-)ethicality. In the theft task, participants are asked to solve some simple mathematical problems within a limited time, grade the answers and pay themselves accordingly. In general, the individual dishonesty level can be measured by

comparing the individuals‘ self-paid amount with the reference point, such as the actual correctly-solved problems or the self-paid amount when individuals cannot cheat.

In this study, the first experiment examined the estimated likelihood of being caught and social norms. There were four experimental conditions in this experiment. First, in the control condition, there was no opportunity to cheat. Second, the shredder condition measured the degree of corrupt behavior in a situation where participants could misreport the outcomes (i.e. the number of correctly solved matrices) with impunity but were not exposed to others‘ corrupt behaviors. The other two conditions elicited artificially-induced in-group and out-group identities of other individuals (specifically, confederates) who explicitly behave corruptly. The second experiment, which aimed to test the independent effect of ethicality saliency, consisted of three conditions: the control condition and the shredder, similar to the first experiment, as well as the saliency condition where confederates enhanced the saliency of cheating but did not change the social norms. The results showed that individuals cheated more when they cannot be caught and, interestingly, the magnitude of individuals‘ corrupt behavior is dependent on group identification. Specifically, the number of reported correct answers was lowest in the control condition and increased in the shredder condition. Moreover, the number of reported correct answers was higher in the in-group condition in comparison to the out-group condition. This suggests that participants behaved more corruptly when observing in-group members cheated but less when observing out-group members‘ unethical behavior. The pattern of the results corroborates that saliency of unethicality and social norms (as well as social identification) both independently play a role in the effect of observing others‘ corrupt behavior. Social norms shape individuals‘ corrupt behavior by providing information about what most people commonly do or which behavior is most appropriate in particular situations (Keizer et al. 2008). On the contrary,

the likely chance of getting caught in corrupt acts does not seem to serve as an explanatory variable for individual actions in the intergroup setting—in contrast to the effect of monitoring found in the individual corrupt behavior studies.

By far, a great deal of research has pushed forward our understanding on individual dishonesty and its relevant psychological mechanisms. However, in certain social settings, corrupt acts cannot be carried out individually, and thus interpersonal cooperation between two or more individuals is required for the success of such endeavors. So, the question arises: What would happen when an interpersonal interaction is a prerequisite for the corruption? Weisel and Shalvi (2015) delved into this issue and experimentally demonstrated that cooperation plays a part in inducing the emergence of corruption in the collaborative setting—especially when incentives are equally aligned. In their experiment, participants played a novel sequential dyadic die-rolling game, where one player confidentially rolls a die and reveals the outcome to his or her partner, who then also secretly rolls a die and records the outcome. At the end of the game, both players receive a payment if, and only if, their reported numbers are the same. For instance, 5 Euros are paid to both players if both of them report 5 from a die-rolling. When the actual results of both players are not aligned, the design of this game creates a dilemma between social and moral sentiments—specifically, between being honest (i.e. reporting the actual misaligned results and, as a consequence, gaining nothing) and lying (i.e. misreporting the results, in order for both players to receive payment). Interestingly, the results showed that the proportion of participants reporting the same outcomes were nearly five times higher than the theoretical baseline of the expected outcome distribution assuming honesty. The authors interpreted the findings as ‗a dark side of cooperation‘. That is, corruption is shaped not merely by individuals‘ personal incentives but also by human tendency to cooperate. From this, another interpretation

could also be that corruption is driven by individuals‘ tendency to comply with a social norm, which in this particular case is cooperation.

On the whole, the bulk of evidence from literatures concerning corrupt behaviors has explored various psychological variables that moderate corrupt behaviors both in the individual and interpersonal settings. Experimental findings have consistently pointed out that individuals‘ corrupt behavior appears to be strongly driven by the intrinsic cost of behaving corruptly, monitoring, and the relevance of moral/social norms. On top of that, when individual incentives equally oriented, interpersonal cooperation as a social norm may be the cause of an increasing magnitude of corrupt behavior in the collaborative setting. Also, importantly, even when the individuals‘ final outcomes do not meaningfully depend on cooperative interactions, people seem to be influenced by others‘ dishonest behaviors. The underlying mechanisms of this phenomenon can be elucidated on the basis of saliency of ethicality and social norms at hand. Nonetheless, in spite of the significance of the role of social norms, which may be to a large extent dictated by the culture-specific conditions, none of the research mentioned above has touched upon how macro-components, e.g. cultural backgrounds, could be a predictor variable on corrupt behaviors. In the next section, connecting the role of culture and corrupt behaviors, I provide an overview of how cultural differences may influence corrupt behaviors.

3. The impact of culture on individuals’ corrupt behaviors

Culture, as defined by Richerson and Boyd (2005), is ‗information that people acquire from others by teaching, imitation, and other forms of social learning. On a scale unknown in any other species, people acquire skills, beliefs, and values from the people around them, and

these variables strongly affect individual behaviors‘. As such, diverse cultural values or backgrounds might render different psychological platforms which in turn drive differences in individuals‘ behavioral patterns. In fact, empirical evidence has provided that there seems to be the effect of cultural difference on individuals‘ corrupt behavior (or at least on the perceived expectation of corruption). For instance, West (2005) examined the return rates of lost wallets in Japan and the US, and found that 85 percent of Japanese citizens returned the lost wallets while only 30 percent of the lost wallets were returned by US citizens. Rosenbaum et al. (2012) empirically found a disparity in fairness and corrupt behaviors between Kazakhstan and Uzbekistan. Robert and Arnab (2013) used the sender-receiver game, which measures the level of dishonesty by looking at individuals‘ tendency to convey false or misdirected information to their counterparts in order to gain financial benefits. The researchers illustrated the effect of cultural difference that the corruption level was higher in India (a more corrupt country) in comparison to the US (a less corrupt country) in accordance with Transparency International's 2005 corruption perception index (CPI). The result seems to suggest that individuals‘ corrupt behavior may be, to some extent, estimated by country‘s level of corruption. Theoretically, this may be explained by social norms. Because social norms describe what most people (from particular groups) commonly do in certain situations, the fact that some countries have high level of corruption may shape the way people perceive corrupt acts (e.g. lying or cheating), which in turn guides their actual behavior. Specifically, individuals from a more corrupt country may think it is commonly acceptable to lie or cheat for self-benefits and thus have a higher tendency to act accordingly.

On top of that, using the die-in-a-cup task, a recent cross-cultural experimental study conducted in 23 different countries has provided empirical evidence on the link between the

prevalence of rule violation and individual intrinsic dishonesty (or, to be precise, the extent to which individuals bend the rules), highlighting cultural backgrounds as a likely causal variable determining the degree of corruption (Gächter & Schulz, 2016). In the experiment, participants roll a six-sided die twice but have to report only the first attempt under complete impunity. This means that there is a zero chance of getting caught and/or punished. Participants earn a monetary reward according to the reported number, except for the number 6 in which the earning would be zero. For example, even if the die is rolled 1 twice but 5 is reported, participants will be paid accordingly. In the end, the overall reported results are then compared to the theoretical baseline distribution of honest reports—for example, theoretically, in a single roll there is equally 16.7 percent probability for each side of a die to be reported. The difference between the two values implies the degree of dishonest behavior at the aggregate level.

In this version of the die-in-a-cup task, reporting the better result of the two rolls indicates ‗justified ethicality‘, which suggests that, instead of brazenly lying, people bend the rules by reporting the highest of the two rolls to maintain an honest self-image (see Shalvi et al., 2011). Furthermore, the researchers also measured an index of the prevalence of rule violation (PRV) at the national level on the countries‘ political fraud, tax evasion as well as corruption and linked it to the behavioral outcomes from the behavioral task. Some of the findings are very fruitful in understanding the role of culture in relation to individuals‘ dishonesty. First, participants from high PRV countries are more dishonest than participants from low PRV countries. Second, the fraction of participants with full honesty (reporting the roll of 6) is also reflected in the PRV; that is, there are more fully honest individuals in the low PRV countries in comparison to the countries with a high level of PRV. In addition, participants from individualist countries, on average, have lower claims on the reported roll than participants from collectivist

countries. This seems to suggest that individuals from individualist societies tend to be less corrupted (Mazar & Aggarwal, 2011). All in all, this study highlights the significance of how culture may influence individuals‘ dishonesty.

Although existing research has established how cultural difference may impact corrupt behaviors, the issue of the effect of intercultural interactions on corrupt behaviors is still unresolved. Consider working in multinational companies or international academic institutions, where talents come from all walks of life across the world. Our main question is of interest in this kind of social setting. Particularly, what would happen to individuals‘ corrupt behavior when individuals of different cultural backgrounds interact with each other? For instance, when a group consisted of German and Italian employees has to work together, will the fact that there is less corruption in Germany, in comparison to Italy, cause a difference and possibly influence changes in their corrupt behaviors, and in what circumstances? Will an Italian working with a German become less corrupted, in comparison to when a German works with their fellow German citizen or with Dutch who has a comparatively low level of corruption? Or, will the German employee behave more corruptly as a result of the mutual work? Or, in a more general sense, would working with individuals from a country which has more/less corruption make one behave more/less corruptly? To my knowledge, no prior research has yet attempted to probe into these questions. Nonetheless, numerous studies in the related domains of behavior (such as cooperation, trust, reciprocity, and fairness consideration) have already explored the issue of how intercultural interactions may affect human behaviors. In the next section, I turn to the research area of economic behaviors and the effect of intercultural interactions by reviewing the behavioral results that may be related to corrupt behaviors.

4. The effect of intercultural interactions on individual behaviors in economic games

Recently, economists have become more interested in the role of culture as a causal variable in economic decisions (Castro, 2008). In light of this, many cross-cultural experiments using economics laboratory games have repeatedly provided evidence suggesting that cultural background is one of the determinants for differences and similarities in individual‘s economic behavior by comparing the results of the same experiment run in different countries or locations (Buchan & Croson, 2004; Cason, Saijo, & Yamato, 2002; Castro, 2008; Henrich et al., 2001; Roth, Prasnikar, Okuno-Fujiwara, & Zamir, 1991).

For example, Buchan and Croson (2004) investigated the relationship between trust along with trustworthiness and social distance, which determines the ‗closeness‘ between parties involved in strategic interactions, in the United States compared with China. The authors detected different levels of trust between the two aforementioned countries. Specifically they found that the level of trust was higher in China than in the US but the level of trust in both countries decreases as social distance increases. Cason et al. (2002) conducted a repeated public good game and showed that American consistently contribute less than Japanese in the last period of experiments. In addition, differences in proposer‘s and/or responder‘s decisions were apparent across cultural groups in bargaining behaviors. Henrich et al. (2001), for instance, demonstrated large variations in the average offer and the rejection rate across 15 small-scale societies. The researchers suggested that these variations are dependent on group-specific conditions such as social institutions or cultural fairness norms, highlighting that different cultures have an influence on differences in individuals‘ behaviors.

In a sense, cross-cultural experiments are intra-cultural experiments done in different countries (Netzer & Sutter, 2009). In other words, cross-cultural studies compare differences between the homogenous within-culture groups of individuals, usually from the same country, interacting with each other. This line of research has been worthwhile in understanding how different cultures may lead to behavioral differences in individual economic decisions. However, these studies do not test how intercultural situations, where individuals from different cultures interact with each other, affect such behaviors. In intercultural interactions, the identity of parties involved (e.g. nationality) is usually known, and consequently, individuals might behave differently when interacting with people who come from different cultural backgrounds. Thus, individual decisions may not only be dictated by their own social norms or values, but also by the perception of their counterpart‘s cultures (Hennig-Schmidt et al., 2007).

In the past few years, research has begun to pay more attention to the impact of intercultural interactions, especially in the context of economic games (e.g., Boarini, Laslier, & Robin, 2009; Carpenter & Cardenas, 2011; Chuah, Hoffmann, Jones, & Williams, 2007; Fershtman & Gneezy, 2001; Castro, 2008; Goerg, Güth, Walkowitz, & Weiland, 2008; Guillen & Ji, 2011; Netzer & Sutter, 2009; Willinger, Keser, Lohmann, & Usunier, 2003). For instance, Fershtman and Gneezy (2001) studied participants in Israeli Jewish society and found a ‗systematic‘ mistrust towards Eastern males, which means both cultural groups trust males of Eastern ethnic origin less. Boarini et al. (2009) studied the norm of equity by comparing bargaining behavior between French and Indian participants. The results reported that unequal splits in favor of French were mostly observed when French participants made a bargain to Indian participants, while almost equal splits occurred most frequently when Indian participants made an offer to French participants. These results seem to suggest that individuals behave

differently in the intercultural interactions where they directly engage with people of different cultures. Hofstede (2001) explained this social phenomenon by suggesting that individuals have different culture-specific characteristics such as power distance and the level of individualism along with their perceived ideas of other cultures that shape their decisions in intercultural interactions. Nevertheless, note that cultural characteristics are not the sole determining factor. Depending on the experimental frameworks, other causal variables are proposed to psychologically drive behaviors at the individual level.

Although a sizeable proportion of intercultural experiments have corroborated the fact that intercultural interactions play a part in moderating individual decisions, not all evidence point to the same direction. Some studies reported no structural differences in participants‘ behavior during the intercultural interactions. For instance, Willinger et al. (2003) compared the economic choices between French and German people and found that although, in the cross-cultural treatment, the average trust level was higher in Germany than in France, difference in the reciprocity level between the two countries did not differ. Also, there was no difference between the intra-cultural and intercultural treatments, meaning no effect of intercultural interactions. Specifically, German trustors did not perceive French trustees as less trustworthy than German trustees and French trustors did not view German trustees as less trustworthy than French trustees, suggesting that both German and French did not differentiate their counterparts on the basis of their country of origin. Similarly, Goerg et al. (2008) demonstrated that, on average, participants from Germany, Israel, and Palestine did not discriminate against their partners based on provenance in the ultimatum bargaining game, neither in the role of proposers nor responders. These results imply that the influence of intercultural interactions is not monotonic and may substantially be a matter of specific behavior and/or relevant contexts.

The important questions, then, arise: What is the effect of intercultural interactions on each specific domain of economic behaviors? In which context does an intercultural interaction affect individuals‘ behavior? Also, what could be the psychological causes or mechanisms that guide the effect of intercultural interactions? Most literatures that investigate the effect of intercultural interactions focus on one specific domain of economic behaviors (e.g. public good contribution, cooperation, trust and reciprocity, or fairness). Unlike in the research investigating unethical behaviors, each domain of behaviors has been studied using the specific economics games (e.g. the public good game, the trust game, the ultimatum bargaining game, or the dictator game). A plethora of different explanations for the observed behavioral patterns have been separately presented in each stream of literatures. From here, I review and examine the results from disparate areas of research pertaining to the impact of intercultural interactions on economic decisions, and discuss the explanatory, psychological causes that have been proposed in the literatures.

4.1 Public good contribution and intergroup cooperation

An important domain of human interpersonal behavior connecting between economics and society is (parochial) cooperation. In general, cooperative behavior in human society may be based on social norms (Fehr & Fischbacher, 2004). However, the issue of cooperation between individuals becomes more complicated when looking at the intergroup setting, especially at the international and global levels. Consider global social dilemma, which describes a global situation in which self-interests and benefits of collective welfare are at conflict—for example, global warming and overharvesting of public commodity. Nowadays, this important matter

requires more research attention, which necessitates the need to understand the issue of large-scale international human cooperation.

Experimentally, a multilevel sequential contribution (MSC) game has been used to explore this issue (see e.g. Buchan et al., 2009). In this game, participants are assigned into groups, which in this case consist of individuals from different countries, and decide whether/how much to allocate to different pools: personal, local, and global accounts. Ultimately, the investment in the global account is the most beneficial if all members of different groups cooperate. In this way, the game design creates a tension between self-interests, parochial interests, and cosmopolitan interests, respectively. Recently, it has been shown that the globalization at the levels of country measured by the country-level globalization index (CGI) and individual both championed cosmopolitan attitudes and systematically predicted individual cooperation at the global and local levels (Buchan et al., 2009). The higher the globalized attitudes; the more participants cooperate at the global level. On top of that, the degree of cooperation is also significantly correlated with the saliency of social or group identity (De Cremer & Van Vugt, 1999). Individuals‘ contribution to a global welfare corresponds strongly with their expectation about other in-group members‘ contribution as well as their self-reported identification with the world as a whole, thereby moderating their personal values relating to global issues (Buchan et al., 2011). These studies support the conjecture that social identification and expectation of what others would do may underlie the effect of intercultural interactions.

Therefore, along these lines, another relevant underlying mechanism that plays a role in intercultural cooperation is intergroup bias or discrimination. In a broader sense, the concept of group can be referred to as ‗a collection of individuals who perceive themselves to be members of the same social category…‘ (Tajfel & Turner, 1979). Tajfel and Turner further identified three

factors that contribute to the emergence of intergroup bias: the degree to which individuals identify with their own groups, the extent to which the prevailing context allows for comparisons across groups, and lastly the perceived relevance of the group comparisons. Thus, in this sense, depending on the social context the idea of group membership can flexibly apply to a wide range of social categories, such as from a small-scale social category (e.g. artificial, experimental groups) to a large-scale social category (e.g. nationality or culture). In particular, groups in intercultural interactions would correspond to the latter case. In this way, the presence of different cultural groups may exacerbate the relevance of intergroup discrimination providing reasons for individuals to differentiate between groups, which underlie the effect of intercultural interactions.

Intergroup bias is closely related to in-group favoritism which suggests that people have a tendency to favor the members who belong to the same group (in-groups) and to discriminate against the members of different groups (out-groups) (for a review see De Dreu, Balliet, & Halevy, 2014). In the stream of literature investigating intergroup cooperation, some studies (e.g. Chuah et al., 2007; Castro, 2008) have proposed in-group favoritism as a driving mechanism underlying the intercultural interaction effect. For instance, using a public good game (similar to the MSC game), Castro (2008) showed that, even in absence of the country effect, a heterogeneous group of UK and Italian participants contribute less to the public pool than a homogeneous group regardless of nationality. This seems to suggest that the social identity of partner has an influence on individuals‘ decision to cooperate. Specifically, when the presence of (cultural) groups becomes salient, the interaction between individuals from different groups may turn out to be less cooperative in the setting where personal and collective interests are in conflict. Accordingly, if indeed individuals expect more cooperation and, in fact, tend to

cooperate more with in-group members (e.g. fellow citizens from the same country or cultural group) than out-group members, and corruption requires cooperation, it might be the case that corruption will be reduced when individuals collaborate with others from different groups.

4.2 Trust and positive reciprocity

In the economic context, nearly all business cooperation and transactions require interpersonal human interactions. One of the main elements in human interactions is trust between individuals—especially in the international trade and financial market. According to Willinger et al. (2003), trust is ‗based on a moral consensus between economic partners, which enables them to stick to their commitments without being obliged to do so by the sole means of formal contracts and legal obligations‘. In some studies, trust is inherently regarded as a primitive code of behavioral conducts—like social norms (sees Fukuyama, 1995). This also means that the level of trust may be translated from the level of individual to that of society or even country, and vice versa. Hence, trust attitude may directly affect performance of institution as well as individual decisions. From this, for instance, Fukuyama makes a contrast between high-trust countries (e.g. Germany and Japan) and low-trust countries (e.g. Italy and France), though only with the proof-by-anecdote methodology. Note that many studies using the trust game to compare between different members of cultural groups intend to shade light on the issue such as ethnic discrimination (Barr et al., 2015; Fershtman & Gneezy, 2001; Guillen & Ji, 2011). However, the results may also be interpreted in general as the ability to distinguish between groups based on stereotypes or cultural backgrounds, which as a consequence influences behaviors during the intercultural interaction.

In the experimental setting, the level of trust and reciprocity can be measured by using attitudinal questionnaires, such as the General Society Survey (GSS)/World Values Survey (WVS), and/or economic games such as the trust game (see Berg, Dickhaut, & McCabe, 1995). In the trust game, one player is called the trustor and the other is assigned the role of trustee. The trustor invests, or entrusts, a certain amount of money (or units) to the trustee. The amount will be increased, e.g. two- or three-fold, and the trustee then decides how much (if any at all) he or she will return, or reciprocate, to the trustor. In this way, the level of trust can be interpreted from the amount of the trustor‘s investment, whereas the level of positive reciprocity can be quantified from the amount given back by the trustee. While a majority of empirical research on trust have been done with a homogeneous group, i.e. without considering participants‘ cultural backgrounds, many previous cross-cultural studies have continually identified differences in the degree of trust between different countries at the individual level (e.g. Buchan, Croson, & Dawes, 2002). Consequently, the discrepancy between different countries‘ levels of trust may have an influence on individual decisions in the intercultural context.

Accordingly, numerous recent studies using the trust game have shaded light on this issue and made important contributions in understanding how interactions between individuals of different cultures play a role on trust and positive reciprocity (Fershtman & Gneezy, 2001; Guillen & Ji, 2011; Hennig-Schmidt et al., 2007; Kim et al., 2013; Netzer & Sutter, 2009; Walkowitz, Oberhammer, & Henning-Schmidt, 2004; Willinger et al., 2003). On the one hand, inconsistent results for the intercultural interaction impact on reciprocity have been reported when comparing between several studies, but not taking in account specific countries (e.g. Hennig-Schmidt et al., 2007; Netzer & Sutter, 2009; Walkowitz, Hennig-Schmidt, & Oberhammer, 2006). On the other hand, the effect of intercultural interactions on trust seems

congruous and some causal psychological determinants for the effect have been proposed in the literatures.

One prevalent driving factor is the perceived cultural characteristics and/or stereotypes of the interacting partners. Importantly, note that the differences in cultural characteristics have not been proposed only in the domain of trust but also in fairness as well as public good contribution (e.g. Carpenter & Cardenas, 2011; Chuah et al., 2007; Castro, 2008). In reality where information is incomplete, individuals may depend on statistical beliefs, e.g. availability and representative heuristics, or a simplified and commonly held perception of people belonging to certain groups, e.g. ethnic, religious and gender stereotypes. Importantly, as stereotypes are usually not one-dimensional, some stereotypes might be statistically accurate while many might be completely fallacious (Fershtman & Gneezy, 2001). For example, Netzer and Sutter (2009) investigated the intercultural trust pattern between Austrian and Japanese participants on the basis of differences in cultural dimensions—one of which is individualism versus collectivism. Japan is considered as a collectivist country, whereas Austria is characterized as an individualist country. The authors showed that Austrian participants trusted Japanese participants more than their fellow citizens; however, Japanese participants did not differentiate between the Austrian and Japanese partners. Furthermore, the post-experimental questionnaires revealed difference in general attitudes towards trust and reciprocity: Austrian participants were more trusting and also viewed themselves as more trustworthy than Japanese participants did. The authors further suggested that the results were in line with the widespread international perception/stereotype of Japanese people. Therefore, in this case, the cultural characteristics seem to influence stereotypes of each nationality, which in turn underlies the effect intercultural interactions have on behavior. On the contrary to previous study of Castro (2008)‘s study on public good contribution, the

effect of intercultural interactions seems to be elucidated on the basis of the perceived cultural identities of group members involved, rather than merely the intergroup bias or in-group favoritism. Hence, during the intercultural interactions, individuals may have more trust in their partners who come from the country with the higher level of preconceived trustworthiness than when compared to even their own fellow citizens. Similarly, in the context of corruption, cultural characteristics and stereotypes of partners might also dictate individual corrupt decisions. For example, when interacting with individuals from collectivist culture, people might turn more corrupt because they think that their partners are likely to cooperate in corrupt acts.

On the opposite side, the fact that individuals may have unequal degree of experiences interacting with foreigners may create different expectations. Individuals‘ beliefs and/or expectations—sometimes false—about their counterparts‘ likely actions have been observed to drive behaviors during intercultural interactions. Interestingly, Hennig-Schmidt et al. (2007) found that participants‘ beliefs clearly correspond to their partners‘ behavior from the same country but are totally different from the actual behaviors of their partners from different countries. Participants‘ beliefs were based on their ethnocentric view, i.e. participants evaluate their counterparts on the basis of standard behaviors of their own culture without considering possible cultural differences. This implies that the prevailing effect of intercultural interactions on individual behavior might arise solely because of the fact that different cultures have different social norms or values—even in the absence of, or independent of, stereotype and prejudice. That is, the resulting behavior during the intercultural interactions is dependent on individuals‘ own level of trust, regardless of their partners‘ social identity. Thus, in this case, individuals from culture with a high level of trust (e.g. Germany) would be more trusting in their partners than individuals from a low trust level country (e.g. Italy). In the similar vein, we might expect

that individuals‘ corrupt behavior would also depend on their own cultural perspective on the standard behaviors (specifically, corruption). For this reason, interactions between individuals from countries with a high level of corruption might induce higher corruption in comparison with interactions between individuals from low corruption countries.

On top of that, how intercultural interactions affects individuals‘ behavior may be determined by social norms (see for a discussion Barr et al., 2015; Henrich et al., 2001; Kim et al., 2013). Substantial evidence from evolutionary psychologists has demonstrated that human beings have evolved the capacity to learn social norms (e.g. Cosmides & Tooby, 1992). Social norms can systematically modify individual actions as individuals mentally identify themselves with certain social or identity groups and behave according to how the groups deem appropriate in certain situations (see Akerlof & Kranton, 2000). Consider arriving for the first time in London, where a majority of people walk in the street without regard for the approaching cars. Though anecdotal, it is likely that many people who are new to the city, especially tourists, will start to jaywalk as well. In this way, social norms may guide individuals‘ behaviors during the intercultural interactions, also in the domain of corruption. For example, when interacting with individuals who obviously and consistently violate rules or engage in dishonest acts, such endeavors may be socially established as a norm and, as a result, might likely increase the level of corruption.

4.3 Fairness and justice consideration

Closely related to the issue of trust and reciprocity in the interpersonal economic interactions is fairness. Though not always, the notion of fairness has been seen as equivalent to

equity, and in some situations equality (Schroeder, Steel, Woodell, & Bembenek, 2003). Nevertheless, the degree of fairness as a social norm is relative; that is, people compare with others as a reference for their fairness consideration (Pillutla & Murnighan, 1996). In an experimental setting, the ultimatum bargaining game (UG) is used to study fairness (or equity) motives as it provides a dilemma between self-interest and the reasons to behave fairly, i.e. for most cases to equally divide the payoffs (see Rilling & Sanfey, 2011). In this game, there are two players: the proposer and the responder—just like in a real bargaining situation. Initially, the proposer is given a certain amount of money and is responsible to decide how the money should be distributed among the two players. The responder, then, can either accept or reject the offer. The payoff is distributed according to the proposal if the offer is accepted. On the contrary, if the responder rejects, neither player receives any payoff. When faced with this kind of dilemma, individuals may consider about the concept of fairness as they choose to cooperate (De Dreu, 2010; Henrich et al., 2001).

In many cases in the ultimatum bargaining game, fairness is understood as equality in the outcome distribution between proposers and responders (De Dreu, 2010), and reciprocity norms following the judgment about fairness has been repeatedly observed in the laboratory setting (Ostrom, 1998). Though, in the intercultural or international setting, marginal value of money differs from country to country—for example, in terms of the average national income and the purchasing power—which could make the assumption of equity more complicated and multi-faceted. Boarini et al. (2009) addressed this matter and extensively discussed three alternative premises for the equity norms: ‗Formal Equity‘, ‗Compensation‘ or ‗Global Equity‘, and ‗Local Equity‘. Formal equity norm follows a standard view that in any case divisible goods (e.g. a certain amount of money) should be equally divided. In contrast, global equity norm, on the one

hand, advocates that players‘ wealth beyond the game should be considered to reduce inequality as much as possible; that is, given a certain quantity of goods the poor should receive more than the rich. On the other hand, local equity regards the equalization of local benefits in a given situation. This means individuals who benefit less from the same quantity of goods should receive relatively more. The results of the ultimatum bargaining game between French and Indian participants supported the notion of local equity, though no ‗beyond-game‘ equity consideration seemed to be taken into account by participants. The results also indicated the effect of intercultural interactions on the rejection decisions. Specifically, Indian participants have a lower rejection threshold than French participants, and this was due to the willingness to compensate for a difference in purchasing powers between the two countries (in some cases known as the currency effect). Importantly, the major implication here is that—like in other domain of behaviors—the effect of intercultural interactions is driven by the social norms (of equity or fairness), which seems to be dependent on country-specific conditions, such as the purchasing power.

In addition to social norms, other conditional factors have also been suggested. One of them is surprisingly the issue of ‗home advantage‘. Sports fans, for an anecdotal example, probably often hear this word all the time during the athletes‘ interview—―Today, we‘ve got our home advantage. And we were able to protect our home floor, thanks to our fans…‖ In the international business deals—e.g. importing or exporting business—this issue might also become more relevant. Not acknowledging and respecting business partners‘ social etiquettes could potentially interfere with the trade agreement. Experimentally, Chuah et al. (2007) reported the impact of home advantage between Malaysian and UK participants playing the ultimatum bargaining game in their home and counterparts‘ countries. Proposers playing on their home land

seemed to offer less than when they played in another country. The authors proposed that this was because of differences in the tolerance level, which is similar to Boarini et al. (2009)‘s the willingness to reject at the different amounts of offer.

However, some research provides contradictory evidence. Goerg et al. (2008), for instance, found that there was no significant, though not entirely negligible, effect of intercultural interactions on proposers‘ beliefs and offers between participants of varying social distances. Neither proposers nor responders systematically conditioned their bargaining decisions based on partners‘ provenance or nationality. Nonetheless, there was still significant decision difference between groups, strengthening the results of other cross-cultural studies on the effect of cultural difference on behavior. In particular, the amount proposed for themselves by participants from Israel was on average more than that of participants from Palestine. This suggests inherent discrepancy in behavior between individuals of different cultural backgrounds, which may underlie the effect of intercultural interactions. On the bright side, research on this topic is relatively new and still progressing; therefore, before making any claims, in the future specific investigations and replication studies might be necessary to further understand the effect of ‗home advantage‘ and the fine-grained influence of intercultural interactions on bargaining behavior.

4.4 Commonalities among the intercultural experiments on individuals‘ economic behaviors

All in all, there may not yet be a straightforward solution to completely untangle the psychological, explanatory causes underlying the effect of intercultural interaction on individuals‘ behaviors. Available empirical evidence, by now, has come from diverse research

frameworks focusing on different topics, notwithstanding has provided fruitful understanding and insights on the issue. Overall, the effect of intercultural interactions seems specific for different economic behaviors and to a large extent depends on various psychological and contextual variables at play. Nevertheless, among the studies pertaining to the effect of intercultural interactions on economic behaviors, some commonalities in the characteristics of the experimental design can be identified and are worth addressed.

First, just like in normal business interactions, in most of the experimental procedures participants‘ affiliations such as nationality and/or academic institutions are usually known. This may cause the psychological process of social identification to be more salient, and hence drives individuals‘ decisions during the intercultural interactions. Second, the intercultural experiment requires participation from different cultural groups, which typically belong to different countries. Due to the limitation of sample pools and their proximity, most experiments have been conducted either via online, quasi-real time procedure (e.g. Buchan et al., 2002; Carpenter & Cardenas, 2011; Castro, 2008; Goerg et al., 2008) or at different points in time across different groups in different locations (Fershtman & Gneezy, 2001; Willinger et al., 2003), in some cases using the strategy method (see, e.g., Hennig-Schmidt et al., 2007; Netzer & Sutter, 2009; Walkowitz et al., 2006). This means that in the intercultural experiments participants usually do not directly interact with each other. Although this might not essentially be problematic for the interpretation of the results in certain settings, it limits the possibility to explore the issue of home advantage during the intercultural interactions.

In the next section, based on the relevant psychological variables and the results from various literatures in the domain of economic decisions, I shall try to postulate and synthesize the

effect of intercultural interactions on corrupt behaviors, specifically in the experimental setting of the novel sequential dyadic die-rolling game from Weisel and Shalvi (2015).

5. A critical synthesis: The effect of intercultural interactions on corrupt behavior

In the review above, I have presented an overview of the experimental literatures pertaining to individuals‘ dishonest behaviors, the effect of culture on such behaviors and, the effect of intercultural interactions on various related behaviors such as public good contribution, cooperation, trust and reciprocity, and fairness consideration. Drawing on accumulated explanations and implications from the existing literatures on economic and corrupt behaviors, I put forward several hypotheses for the effect of intercultural interactions on corruption (as defined in the second section), particularly in the sequential dyadic die-rolling game.

As a reminder, the game involves two players. The first player rolls a die in private and reports the outcome to the second player, who subsequently also privately rolls a die and reveals the resulting outcome. Both players earn a payment equal to the reported outcomes if, and only if, the outcomes from both players are the same (i.e. a double). From here on, I shall refer to the first player as player A and the second player as player B. For player A, consistently reporting the maximum number of 6 would indicate ‗brazen lying‘, while for player B consistently reporting doubles would be considered as brazen lying. Similar to other intercultural interaction experiments, in our case of interest, each player should come from different cultural groups. Though, I acknowledge that other psychological underlying mechanisms might be more relevant in different tasks and/or different contexts. In this article, I specifically speculate the effect of intercultural interactions in relation to the main experiments and findings from the study of

Weisel and Shalvi, namely the overall proportion of corrupt behavior and behaviors of player A and player B in a situation where the outcomes are aligned. In my postulations, the countries‘ levels of corruption as well as individualism are based on the data from Transparency International‘s 2015 corruption perception index (CPI) and the extended Data Table 1 from the study of Gächter and Schulz (2016).

Unlike in a one-shot task where decision-making may be heavily reliant on incomplete information at hand (such as nationality of the partner), in the sequential dyadic die-rolling game there is an opportunity for participants to continually observe their partners‘ actual behavior after the previous attempts. This may make the role of psychological mechanisms such as judgment based on stereotypes or reliance on ethnocentric views less relevant because individuals could change or adjust their decisions after the first few rounds of having observed or learned about their partners‘ behaviors. In addition, because the game design prompts more cooperative action and provides no financial conflict between personal and collective interests (in contrast to, for instance, a public good contribution game), intergroup bias also unlikely plays a prominent role on corrupt behavior in this particular context. Therefore, here, I focus more on other likely explanatory variables including the country‘s levels of corruption and trust, which both may define particular social norms of different cultures, and the cultural dimension of individualism versus collectivism.

First of all, since, it seems, in general people have a tendency to comply with a social norm of cooperation on equal terms, even in the sample pools from Germany and UK, where there is considerably low corruption (Weisel & Shalvi, 2015). From this, we might expect that overall collaboration will turn people to be more corrupt when the incentives are aligned, in line with the result from Weisel and Shalvi‘s study.

H1A: Overall, the proportion of the reported better outcomes will be higher than the proportion according to the theoretical baseline assuming honesty in the aligned outcomes, regardless of cultural groups.

Nonetheless, the degree of collaborative corruption may depend on the countries or cultures that players participating in the experiment belong to. According to previous research, there seems to be difference in corrupt behavior on the basis of countries‘ corruption level (e.g. Gächter & Schulz, 2016). The level of country‘s corruption may reflect social norms which particularly determine the generalized acceptability or appropriateness of corrupt behavior for each country, thereby shaping the (expected) magnitude of individual corruption. Hence, it is likely that the corruption level may depend on the paring between interacting individuals who are driven by specific social norms they adhere to.

H1B: The overall degree of collaborative corruption will be lower for dyads which both players are from the low corruption countries (e.g. German and Dutch participants), than the interaction between individuals from the low- and high- corruption countries (e.g. German and Italian participants). Dyads in which both players are from the high corruption countries will exhibit the highest corruption level (e.g. Chinese and Italian participants). Moreover, aside from the country‘s corruption level, since there is no conflict between personal and collective/group interests, it is unlikely that there will be the effect of intergroup bias. In other words, as long as the countries‘ corruptions are relatively at the same level, there will be no behavioral difference in corruption, for example, when comparing the interactions between German and Dutch individuals with the interactions between two Dutch individuals.

Next is the corrupt behavior of each player in the intercultural interaction. Interacting with individuals from different cultures may prompt change in individuals‘ corrupt behavior; player A‘s decisions might strategically influence player B‘s decisions, and vice versa. Since in the sequential dyadic die-rolling game the influence of each player‘s decision on the outcome is not symmetrical, there might be variations in the number of players A and B who engage in corruption.

H1C: The magnitude of likelihood of corruption, especially brazen lying (i.e. always reporting doubles), will be higher when player B is from a country with high level of corruption compared to a country with low level of corruption.

On the other hand, one similar element among the first player (i.e. player A) in the sequential dyadic die-rolling game, the trustor in the trust game and the proposer in the ultimatum bargaining game is that they initiate or set the tone of the interaction by making the first decision. This is important because culture-specific variables, such as the level of corruption, may inherently dictate individuals‘ decisions in the beginning period of the game, and possibly continually so throughout the experiment.

H1D: The magnitude of likelihood of brazen corrupt behavior, i.e. always reporting the roll of 6, will be lower when player A is from a country with low level of corruption compared to a country with high level of corruption.

Additionally, in a situation where individuals (both as players A and B) interact with a consistent brazen liar counterpart, the observed corrupt behavior might be recognized as a contextual norm (see Keizer et al. 2008), that is, to lie about the outcome for the final collective benefit. This may consequently influence individuals‘ behavior as collaborative corruption is

ostensibly highlighted as a social norm. From this, it is likely that the number of individuals who will turn corrupt may also be related to the country‘s level of corruption.

H1E: There will be a positive relationship between the number of individuals who turn corrupt and the country‘s level of corruption. That is, the more corrupt the country; the more number of individuals who turn corrupt.

Next determining variable is the country‘s level of trust. In light of this, research has suggested a positive relationship between trust and honesty (e.g. Mann et al., 2016). Specifically, individuals who have more trust in others behave more honestly. Therefore, as established earlier that there is the effect of intercultural interactions on individuals‘ trust and consequently their decisions, we might expect a similar trend for corrupt behavior in relation to the country‘s level of trust.

H2A: The magnitude of corruption will be higher in case the interaction involves individuals (both as player A and B) from low trust cultures in comparison to high trust cultures.

On the opposite side, trust between individuals is required to engage in corruption. For example, because trust can be developed in a long-term relationship, Abbink (2004) demonstrated that the rotation between interacting partners reduces corruption in comparison to the condition where the pairing of partners is fixed. As such, individuals who are more trusting may develop trust more easily and hence are more likely to engage in corruption.

H2B: The magnitude of corruption will be higher when the interaction involves individuals (both as player A and B) from culture with the high level of trust in comparison to low trust cultures.

Finally, another explanatory factor is cultural dimensions, particularly the degree of individualism, which may determine the intrinsic psychological cost of lying. The game creates a psychological tension between a propensity to maintain an honest self-image and a tendency to be cooperative (positive self-conception maintenance versus cooperation), especially for player B. Since collaboration on equal terms seems to reduce a psychological cost associated with lying (Weisel & Shalvi, 2015), people from individualist countries may likely suffer more from the psychological cost of lying in order to cooperate, while the cost of lying for people from collectivist countries may be intrinsically mitigated by an act of successful cooperation. Hence, in accordance with previous research in the individual setting (Gächter & Schulz, 2016; Mazar & Aggarwal, 2011), we might expect the magnitude of corruption to be related with the degree of individualism.

H3A: The average magnitude of corrupt behavior will be higher in the interactions between individuals from collectivist countries (e.g. Japan and China) than from individualist countries (e.g. Germany and The Netherlands).

Importantly, in the current setting the cost of corrupt behavior, which is lying about the outcome, is not equal for both players since the psychological burden of corrupt behavior substantially presses on player B (Weisel & Shalvi, 2015). Hence, the degree of individualism may determine the likelihood that individuals will engage in corruption in order to achieve the final collective outcomes.

H3B: Player B from collectivist cultures will behave more corruptly (i.e. correspond to more brazen lies) than player B from individualist cultures.

In addition, previous research has shown that the degree of individualism influences individuals‘ willingness to engage in bribery (Mazar & Aggarwal, 2011). Specifically, the higher the degree of collectivism; the more likely individuals are willing to offer a bribe (or, more broadly, to initiate corruption). Therefore, the degree of individualism may determine the likelihood that individuals will initiate corruption, which is to always report the roll of 6.

H3C: The magnitude of brazen corrupt behavior, i.e. always reporting the roll of 6, of player A will be higher for individuals from the country with a low degree of individualism compared to the country with a low degree of individualism.

To investigate these effects at the individual level, future research may consider increasing the number of repeated trials for each dyad so that the observed behavior of counterparts becomes more prevalent.

6. Conclusion and future research

The current review focuses on connecting two important research issues: the role of culture (i.e., more specifically, the intercultural interaction) and corrupt behavior. In this review, I highlight the importance of understanding the effect of intercultural interactions on corrupt behaviors. I provide an overview of empirical evidence on individual dishonest behaviors and psychological variables that potentially underlie such behaviors on particular contexts, such as intrinsic psychological cost, social norms, saliency of ethicality, and presence of the third party. I, further, discuss the relevance of culture and cultural differences on corrupt behaviors. Then, I review the effect of intercultural interactions on economic behaviors—namely, cooperation, trust and reciprocity, and fairness consideration—and identify some common explanatory