The two issues that impact compensation practices in an international business are

The compensation gap between locals and expatriates may be viewed as the result of an extreme form of a two-tier wage system. Traditionally, a two-tier wage system refers to the arrangement in which new employees are put under a compensation scheme that is substantially lower than the scheme for existing employees (Lee & Martin, 1991; Martin & Peterson, 1987). The two-tier system for locals and expatriates in a foreign operation of a multinational differs from traditional two-tier systems in that the categorization criterion is based on country of origin and cultural identity. Furthermore, the differentiation between the two tiers tends to be much larger, which reflects the difference between the levels of economic development of the countries involved. In developing countries, it is common that the salary of expatriates is several times of that of locals holding similar positions.

A justice perspective is able to shed light on this extreme form of a two-tier pay system. Equity theory posits that people compare their rewards with similar others. Fairness is perceived when the ratio of inputs and outcomes are similar across individuals (Adams, 1965). Based on equity theory, Lawler (1981) proposed a pay satisfaction model, in which perceived pay of referent others is able to influence pay satisfaction. In the Chinese context, locals who work for multinational firms are typically paid higher than their counterparts in local firms, especially those who work in state-owned enterprises, but their pay is only a small fraction of that of expatriates. If locals compare with employees of state-owned enterprises, they are unlikely to experience a sense of injustice. If they compare with expatriates, however, research on relative deprivation suggests that they are likely to see themselves as victims of gross injustice (e.g., Sweeney, McFarlin, & Inderrieden, 1990; Tougas, Rinfret, Beaton, & de la Sablonnière, 2005). Previous research has shown that locals indeed compare with expatriates and report a high level of distributive injustice (e.g., Chen et al., 2002; Leung, Wang, & Smith, 2001).

It is difficult to raise the salary of locals drastically to reduce the injustice associated with the compensation disparity between them and the expatriates. Justice theories and research are able to point to some directions for attenuating the injustice associated with the compensation disparity in multinationals in developing nations. Although justice effects are well documented, i.e., perceived injustice is related to a wide range of negative reactions, a stream of research has examined a variety of moderator variables that are able to reduce the negative impact of injustice. In the absence of similar research in the MNC context, we need to extrapolate from this research to identify factors that are able to mitigate the negative impact of the injustice associated with the compensation disparity. A good starting point is to focus on well-documented effects in the justice literature.

Two types of variables have received a great deal of attention in justice research. Relational variables are concerned with the relationship between social actors, probably because of the prominence of group value model (Lind & Tyler, 1988) and the relational model of authority (Tyler & Lind, 1992). It is well established that the trustworthiness of an authority plays a moderating role in justice perception. If an authority is seen as trustworthy, perceived procedural fairness shows a weaker effect on people's reactions (van den Bos, van Schie, & Colenberg, 2002; van den Bos, Wilke, & Lind, 1998). Outcome-related variables, such as economic gains and promotion, have also received considerable attention in justice research. Relevant to our analysis, Brockner and Wiesenfeld (1996) concluded after a thorough literature review that the effects of procedural justice are smaller when outcome favorability is higher then when outcomes are unfavorable. In other words, if one receives a better outcome, one is less likely to react negatively to procedural injustice. This stream of research is concerned with variables that moderate the effects of procedural justice. In our context of interest, however, the perceived fairness of the compensations of locals and expatriates is involved. The key issue is distributive rather than procedural justice. Nonetheless, this line of work provides the basis for a sound theoretical analysis of the issues involved.

In summary, we focused on two variables that are widely known to be able to moderate the effects of procedural justice in previous studies: trustworthiness of expatriates and perceived compensation. Trustworthiness is relational in nature, whereas perceived compensation involves tangible gains. The extensive justice literature with regard to these two variables allows us to develop some well-grounded hypotheses. In the following sections, we first describe the two moderator variables of the study, and the outcome variables studied are described in a subsequent section.

Chen et al. (2002) pioneered the research on the moderators of the negative effects of the compensation disparity between locals and expatriates. Specifically, in a survey of local employees of foreign multinationals in China, Chen et al. (2002) found that if expatriates were seen as interpersonally sensitive, such as being kind, helpful, and respectful of locals, the negative impact of the compensation disparity on locals’ perceived compensation fairness was mitigated. In our research, we followed the orientation of mainstream organizational justice research by focusing on variables that moderate the impact of perceived injustice on outcome variables. Nonetheless, Chen et al.'s findings point to the role of a positive relationship between locals and expatriates as a cushion for the negative effects of the injustice associated with the compensation disparity. Given that trustworthiness is probably one of the most important definers of a positive interpersonal relationship (e.g., Lewicki & Bunker, 1995), a reasonable extrapolation is that trustworthiness of expatriate managers is a likely candidate for moderating the negative impact of perceived distributive injustice associated with expatriates.

The fairness heuristic theory provides a theoretical justification for the moderating role of trustworthiness on justice effects (Lind, 2001; van den Bos, Lind, & Wilke, 2001). The central argument of this framework is that one's relationship with authority figures and the favorability of the outcome received from them are often ambiguous and hard to judge. Thus, people need to ascertain that the authority figures have not taken advantage of them, and the fairness treatment received from authority figures constitutes an important cue for this judgment. An important derivation of this reasoning is that if there are other cues that signal the lack of an exploitative intent on the part of authority figures, the role of justice is likely to diminish. As predicted, van den Bos and his colleagues showed, in a series of ingenious studies, that trustworthiness is able to mitigate the effects of procedural justice. If an authority is seen as trustworthy, people's reactions are less affected by perceived procedural fairness (van den Bos et al., 2002, van den Bos et al., 1998). We expect that the fairness heuristic argument should extend to distributive justice, and that trustworthiness is able to buffer the negative impact of distributive injustice on outcome variables. In a multinational context, if local employees report more trustworthiness of expatriates, the negative effects of perceived distributive injustice associated with the compensation gap between locals and expatriates should be attenuated.

It is well known that compensation shows wide-ranging positive effects on job attitudes and reactions (e.g., Sarata, 1984). In procedural justice research, positive outcomes can elevate the perception of procedural justice, generally known as outcome effects in the justice literature (Lind & Tyler, 1988). More relevant to our study and as mentioned before, Brockner and Wiesenfeld (1996) concluded after a thorough literature review that an outcome is able to moderate the effects of procedural justice. If an outcome received is positive, the negative effects of perceived procedural injustice on a wide range of dependent variables are weakened. The outcome × procedural justice interaction is very robust, and we conjecture that it should extend to distributive justice. Following the justice literature, we used the perceived favorability of the compensation received by locals to represent their outcomes, and we expect that perceived compensation should be able to buffer the negative effects of perceived distributive injustice. In other words, if locals receive more favorable compensation, they should react less negatively to the perceived injustice of their salary vis-à-vis that of expatriates. This prediction can be regarded as an extension of the outcome × procedural justice interaction identified by Brockner and Wiesenfeld (1996) to the realm of distributive justice in a multinational setting.

Before formal hypotheses are proposed, the specificity of justice effects needs to be taken into account. Although justice effects are pervasive, previous research has clearly documented the specificity of these effects with regard to the source of the justice treatment. In a review of the literature, Cropanzano, Byrne, Bobocel, & Rupp (2001) proposed a model to describe the source effects in judgments of fairness, which distinguish between the organization and supervisors. In general, people react more strongly to variables that are closely related to the source of their fair or unfair treatment. Justice treatment attributable to the organization shows stronger impact on dependent variables associated with the organization, such as organizational commitment. In contrast, justice treatment attributable to supervisors shows stronger impact on dependent variables associated with supervisors. For instance, interactional justice received from supervisors is related to supervisory commitment, whereas procedural justice (measured by perceived fairness of formal procedures) is related to organizational commitment (for a review, see Bobocel & Holmvall, 2001).

In the context of the operations of MNCs in China, we make a distinction between holistic reactions that are tied to the organization, including organizational commitment and job satisfaction, and specific reactions toward expatriates, including evaluation of and satisfaction with them. Organizational commitment is an organization-oriented attitude. Job satisfaction is determined by many factors, such as pay and promotion. We argue that job satisfaction is more closely tied to the organization than to one's supervisors because there seems to be more determinants of job satisfaction that are attributable to the organization, such working conditions and compensation terms. The source specificity argument suggests that the moderating effect of trustworthiness of expatriates should be stronger for expatriate-related dependent variables, including evaluation of and satisfaction with expatriates. The rationale for this assertion is obvious because the trust is directed toward expatriates, not the entire organization. In contrast, the moderating effects of perceived compensation should be stronger on dependent variables that are tied to the organization, including organizational commitment and job satisfaction. Compensation decisions are more influenced by organizational policies and rules than by the preferences of individual expatriates, and the reactions to this variable should be holistic. The two major hypotheses of the study are now stated as follows: