Answer questions

ARTICLE IN PRESS

international

business

review

International Business Review 16 (2007) 377–395

0969-5931/$ -

doi:10.1016/j

� Correspo

E-mail ad

www.elsevier.com/locate/ibusrev

New evidence in an old debate: Investigating the relationship between HR satisfaction and turnover

Shaista E. Khilji a,�

, Xiaoyun Wang b

a The George Washington University, Washington NW, DC 20052, USA

b Asper School of Business, University of Manitoba, Canada

Received 9 May 2006; received in revised form 28 September 2006, 22 November 2006, 19 December 2006;

accepted 21 December 2006

Abstract

The debate relating to an interconnection between organizational performance and HRM may be

old, but the evidence presented in present analysis is new and reflects the critical value of employee

satisfaction with HR practices, referred to as HRS. Turnover is chosen as a dependent variable

because of its centrality in improving organizational performance. In order to advance the scope of

employees’ day-to-day experiences with HRM, a multiple respondent research strategy is adopted.

Pakistan, with its current socio-economic transitions, presents a rich context in which to study HRS

and turnover. Several demographic and organizational factors of turnover are also included as

moderators. Findings, based upon an analysis of 508 questionnaires in 12 organizations, reveal HRS

is negatively related to turnover. Employee age and organizational performance are also found to

moderate this relationship significantly. Explanations of results reveal relevance of a contextual

paradigm in strategic human resource management and highlight prominence of HRS in this debate.

r 2007 Elsevier Ltd. All rights reserved.

Keywords: Developing country; Human resource management; Pakistan; Performance; Satisfaction; Strategic

human resource management; Turnover

1. Introduction

The strategic human resource management (SHRM) debate, with its proposition that effective human resource (HR) practices positively relate to enhanced employee motivation

see front matter r 2007 Elsevier Ltd. All rights reserved.

.ibusrev.2006.12.005

nding author. Tel.: +1 202 994 1146.

dresses: [email protected] (S.E. Khilji), [email protected] (X. Wang).

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395378

and commitment, is widespread (Datta, Guthrie, & Wright, 2005; Huselid, 1995). However, in a majority of studies attention has been paid to developing and/or testing a comprehensive list of best practices (Arthur, 1994; Lawler, 1986; Pfeffer & Veiga, 1999). Such a tradition, with its emphasis on applying general rules universally, has been facilitated by single respondent data, usually a senior manager. This is surprising since it is individual employee satisfaction with HR practices that fundamentally contributes to a positive relationship between human resource management (HRM) and organizational performance (Walton, 1985). Hence, focus upon capturing presence or absence of certain best practices to the exclusion of addressing employee satisfaction constitutes a serious gap in the existing literature (Khilji & Wang, 2006). Many authors have argued that overlooking non-managers’ and middle managers’ experiences with HRM is likely to lead to skewed findings (Keenoy, 1999; Khilji & Wang, 2006; Lau & Ngo, 2004; Purcell, 1999; Truss, 2001). In order to address this issue, and to inject new evidence to the existing SHRM debate, present study directly evaluates employee satisfaction with various HR practices (hereafter referred to as HR satisfaction or HRS), using multiple responses from each sample organization. Over the years, it has been observed that management often fails at HR implementation,

and that several HR policies eloquently described by senior managers during interviews exist only on paper (Khilji & Wang, 2006; Storey, 1992). The gap between management policies and shop floor reality is related to employee dissatisfaction. This evidence has led some scholars to question the effectiveness of implementation within organizations. Gratton and Truss (2003) elevated the relevance of implementation by including it in their three dimensional people strategy. They argue the mere existence of HR policies on paper is not sufficient, and implementation (i.e. practice or translation of HR policies into action) is fundamental to the question of whether or not an organization is delivering in the area of people management. HRM must take effect in the day-to-day experiences of employees to result in their satisfaction. ‘A key message is that bridging from business goals to employee performance requires not only policies but also a determination to act, as seen through actual practices’ (Gratton & Truss, 2003, p. 75). Thus, for the purpose of present analysis, HRS is defined as employee satisfaction with actual and implemented HR practices. Although HRS construct may appear to have some similarities to the well-researched concept of job satisfaction on the basis of their assessment of job-related dimensions, we believe HRS differs from job satisfaction. The content of job satisfaction is exclusively affective (Mitchell, Holton, Lee, Sablynski, & Erez, 2001) as it is often thought of an ‘emotional state resulting from the evaluation of one’s job experiences’ (Harrison, Newman, & Roth, 2006, p. 306), and its target is one’s job position. On other hand, HRS evaluates both cognitive and affective degree of satisfaction and it targets the entire organization in terms of its focus on HR policies developed and implemented by the upper management. Since the early work of March and Simon (1958), who conceptualized employee

turnover intention as a reflection of an employee’s decision to participate and perform in the organization, turnover has been used as an organizational performance indicator. For most part, voluntary turnover is treated as a managerial problem that requires attention, thus its theory has the premise that people leave if they are unhappy with their jobs and job alternatives are available (Hom & Kinicki, 2001). Several scholars have argued that a high turnover rate clearly has a negative impact on an organization (Phillips, 1996; Pfeffer, 2005). In specific, Phillips (1996, p. 180) emphasizes the centrality of voluntary turnover in

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395 379

SHRM debate by stating, ‘HR programs designed to reduce turnover can result in tremendous bottom line improvements’. With this view in mind and in an effort to develop a meaningful understanding of HRS we decided to include turnover intention as an indicator of organizational performance in the present analysis.

SHRM studies frequently employ the resource-based view to frame research examining links between HRM and organizational performance. For example, Wright and McMahan (1992, p. 298) define SHRM as ‘‘the patterns of planned HR deployments and activities intended to enable an organization to achieve its goals’’, and Wright and Snell (1991, p. 204) state SHRM is ‘‘HR activities used to support the firm’s competitive strategy’’. Such explanations clearly imply employees are value creating assets and a source of unique competitive advantage for organizations, thus it can be argued that their retention (measured in terms of turnover rates) is central to studies focusing upon effectiveness of HR practices. Hence investigating a relationship between turnover and HRS within the framework of SHRM, as is being targeted in present analysis, is well-placed in SHRM literature.

Like other management issues, there is a dearth of turnover and SHRM research in developing countries despite calls to widen the international HR perspective (Baruch & Budhwar, 2006) and concerns with boundary conditions pertaining to SHRM theory (Wright, Scott, Snell, & Dyer, 2005). Consequently, HR practitioners there face a lack of focused strategies for combating high turnover rates and improving organizational performance (Khatri, Budhwar, & Fern, 2005). Developing countries’ recent surge of FDI inflow to reach the US$ 233 billion mark (UNCTAD, 2005), indicates a vibrant business environment. Large scale economic activity also provides impetus to business growth. At the same time this activity pushes managers toward developing new practices, aiding in the development of competitive advantages and the retention of employees who now have many alternative jobs available to them. These challenges are quite apparent in Pakistan, whose recently released economic growth rate of 8.4% makes it the third fastest growing economy in the world (Economist, 2006; Government of Pakistan, 2005). As expected, this rapid pace of economic development has precipitated a change in Pakistani culture, evidenced by the introduction of several western HR practices, including flextime, empowerment, open discussions, MBO, flatter structures, and performance-based remuneration (Khilji, 2004; Mitchell, 2005). Furthermore, analysis of Pakistani employees exhibits younger employees, influenced by American management education at an earlier age in schools and colleges are significantly more inclined towards accepting fair, open, and egalitarian HR systems at workplace when compared to older employees (Khilji, 2004). A similar pattern of changing expectations has also been observed in other fast developing countries, such as India and China. For example, Warner (2004) argues that Chinese recruited from prestigious business programs are comfortable with modern HRM, and follow best practices in SHRM, as exemplified in local firms and multinationals. Chatterjee and Pearson (2000) in their study of Indian managers also demonstrate the emergence of a duality of values among them, and highlight the need for organizations to balance tradition and change during times of rapid economic development. We were therefore motivated by an assortment of social, cultural and economic developments in Pakistan, and believe that a study of employees in an environment undergoing transitions will broaden our understanding of turnover and HRS issues and contribute to the generalizability debate of international SHRM theory. This is particularly important since one of the major issues

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395380

facing the field of SHRM is external validity and generalizability of theory (Wright et al., 2005). The present research will therefore fulfill three main objectives. First, it will emphasize

the prominence of HRS as a primary variable in SHRM debate. Empirically this has not been achieved before. Second, the present study will explore the relationship between HRS and voluntary turnover in light of resource-based view and centrality of turnover in organizational performance (Phillips, 1996; Pfeffer, 2005). Some non-job moderators will also be included in our analysis in order to develop a more comprehensive understanding of turnover. Finally, since a majority of SHRM studies have been conducted in developed countries, the present research will contribute to the literature by drawing its sample from Pakistan, where there is evidence of a fast-paced economic change and precipitating shifts in employee attitude towards HRM. The paper is organized in four main sections. We will begin by exploring the turnover

theory to enumerate several research hypotheses. This will be followed by a discussion of our sample and research design. Results will be presented next. We will conclude with a synthesis of the main findings, limitations and a discussion of implications for future research in SHRM.

2. Theoretical background and hypotheses

Although it has been suggested that models of voluntary turnover are well established (Iverson & Pullman, 2000; Hom & Griffeth, 1995), we argue these are founded upon data sets collected primarily in USA or other developed countries. Therefore, on the basis of a divergence perspective in international HRM and the current evidence from turnover literature which indicates diverse groups of employees exhibit different tendencies to leave (Hom & Kinicki, 2001; Iverson & Pullman, 2000), we question applicability of established models in an emerging and culturally distinct economy, such as Pakistan. We believe that present analysis of employees in Pakistan will contribute to the existing turnover models, in terms of its focus upon data from a developing country. However, at this stage, in the absence of a collection of evidence from Pakistan and other developing countries, we rely upon the broad nature of existing turnover models in our analysis. This includes incorporating a discussion of moderating variables (such as age and gender) and using explanations from existing models as starting points, which will be combined with cultural- specific information from Pakistan to develop hypotheses for present research.

2.1. Research hypotheses

Previous studies indicate that voluntary turnover intention is influenced by the organization’s investment in commitment-enhancing HRM systems (Arthur, 1994; Shaw, Delery, Jenkins, & Gupta, 1998). Research frequently refers to the notion that employees focus upon their own financial and psychological outcomes and that they remain with organizations when staying maximizes self-interest. It is also believed that satisfied employees tend to work harder at their jobs in order to achieve organization’s objectives. Their positive attitude is positively associated with motivation and expressions of loyalty, and negatively related to potentially costly behaviors such as voluntary turnover (Hom & Kinicki, 2001). Hom, Caranikas-Walker, Prussia, and Griffeth (1992) suggest a linear and rational turnover decision sequence whereby employees become dissatisfied with their jobs,

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395 381

form intentions about quitting, which are followed by actual quitting. Maertz (2001) refers to the affective force as a primary cause of turnover. He explains that employees are either comfortable or uncomfortable with the management in their organizations. Comfort relates to satisfaction and discomfort, on the contrary results in the desire to quit. Since HRS is an attitudinal assessment of HR practices developed and implemented by the upper level management, it directly signals employees’ comfort with the management in their organizations. In addition Walton (1985), in specific, draws on a positive relationship between good HR practices, employee satisfaction and their willingness to contribute to organization’s success. Together these arguments lead us to conclude that higher levels of HRS decrease the likelihood of voluntary turnover intentions. In Pakistan, as mentioned previously, economic changes have helped create new jobs. At the same time, socio- cultural changes reflected in the adoption of new HR systems as well as the success of several new western HR practices (such as performance-based rewards, flex time etc.), indicate that we can expect the same HRS-turnover intention relationship to hold true for our sample. It leads us to propose the following hypothesis:

Hypothesis 1. HRS is negatively associated with voluntary turnover intentions.

Maertz (2001) and March and Simon (1958) argue that severing the relationship with the organization entails more than dissatisfaction-induced leaving and can include several non-job factors such as age, gender or availability of attractive job opportunities elsewhere. In addition, literature shows that multinationals in developing countries are regarded as effective vehicles for transferring new management practices (UNCTAD, 2005). This aspect, as will be discussed later, is likely to have an impact on employees’ turnover intentions. Similarly, a category of high performing and low performing organizations is of paramount importance in SHRM literature since the promise of competitive advantage (or high performance) is extended to those organizations that incorporate SHRM themes. We therefore include both individual demographic and organizational factors as moderating variables to help us determine the relationship between HRS and employee turnover. A review of individual and organizational variables is important in view of recent efforts to break away from the traditional attitudes and alternative models prescribed by early turnover writings (Mitchell et al., 2001).

Below is a discussion of various non-job individual demographic and organizational factors, which have been shown to relate moderately to strongly to turnover intentions in the existing turnover literature.

Age: Lee and Mitchell’s (1994) ‘unfolding model of voluntary turnover’ and its extension (Lee, Mitchell, Holtom, McDaniel, & Hill, 1999) includes the notion of lack of fit with employees’ values, implying if some aspect of a job is inconsistent with an employee’s values or goals, it leads to voluntary considerations to quit. Existing evidence from China and India suggests changes in younger generations’ values have been guided by extensive globalization (Chatterjee & Pearson, 2000; Cooke, 2005; Fang, 2006; Herbert, Chatterjee, & Heuer, 2006; Warner, 2004). The same is true in Pakistan where exposure to American management education has spearheaded adoption of open, egalitarian and empowerment oriented HRM systems in several multinationals and private sector organizations (Khilji, 2004; Mitchell, 2005). Thus, we can expect changes (in HR practices) to be more in line with younger employees’ values and goals. This is likely to lead to higher HRS and lower turnover among younger employees, when compared to older employees. At the same time, if younger employees are not satisfied with HR practices, they will quit jobs for

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395382

organizations whose HR practices are more aligned with their values and expectations. We therefore hypothesize:

Hypothesis 2. Voluntary turnover intentions among younger employees are more likely to be

negatively influenced by an increased HRS.

Gender: As mentioned previously, findings of the studies on the relationship between gender and turnover are mixed. Cotton and Tuttle (1986), Weisberg and Kirschenbaum (1993) found females are more likely to quit than males. Campbell and Campbell (2003) and Wai and Robinson (1998) reported no moderating effect of gender on turnover. Recently, Elaine (1997) and Khatri et al. (2005) found males have significantly higher turnover intentions. Such contradictory findings make predicting the effect of gender rather difficult. However, March and Simon’s (1958) argument that the voluntary turnover decision involves community and family (and can trigger leaving) appears to make sense in a collectivist country like Pakistan, where individuals live within protective networks of extended families with traditionally pre-defined gender roles. Although women in Pakistan are increasingly participating in the workforce, their primary job is to raise children and to look after extended families. They are therefore under greater pressures to fulfill domestic roles and may pay less attention to their jobs, careers, and HR systems in their organizations. It implies, irrespective of the level of HRS, females are less likely to stay at their jobs. On other hand, males are more likely to be more influenced by positive and negative HRM systems since they enter the workforce to grow and to develop their potential. Thus, a higher or lower satisfaction with HRM is likely to impact men’s intention to stay or to leave. This leads us to hypothesize:

Hypothesis 3. Voluntary turnover intentions among men are more likely to be negatively

influenced by an increased HRS.

Organizational type—multinational or local: It is generally believed that in developing countries, multinational companies have better HRM systems when compared to local companies (Duarte, 2001; Taylor, Beechler, & Napier, 1996). For example in China and India, multinationals are considered to be ‘‘pattern-setting employers’’; these are gradually leading to the emergence of global value paradigms among their employees (Chatterjee & Pearson, 2000; Warner, 2004; Zhu & Warner, 2000). At the same time Warner (2004) argues that local organizations are still very far from applying the concept of HRM as is understood internationally. Previous studies from Pakistan have also indicated global companies are regarded as training grounds for local employees and that local companies (in general) copy HRM systems from their global counterparts (Khilji, 2004). Additionally multinationals recruit and retain the top talent by offering better career opportunities, compensation and less hierarchical HR systems. Since the top talent is either recruited out of the best business schools in Pakistan (which follow American management curricula) or has been educated in the West, multinational employees are more comfortable with western management philosophy. Socializing and working in multinationals further changes employees’ work-related values to reflect the parent company’s culture. Therefore in a competitive and growing economy where job opportunities are numerous, employees in multinationals (with greater exposure to foreign management philosophy when compared to employees in local organizations) are more likely to be influenced by HR practices in their organizations. Thus, high or low HRS levels are expected to affect their

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395 383

turnover decisions more than employees in the local organizations. This rationale leads us to hypothesize:

Hypothesis 4. Voluntary turnover intentions among employees of multinationals are more

likely to be negatively influenced by an increased HRS.

Organizational type—high performing or low performing: Evidence from the UK and the USA also illustrates that profitable organizations are more likely to invest in developing good HR systems and those undergoing financial crisis are more likely to lay off employees and adopt other cost cutting measures (Lawler, 2005; Pfeffer, 2005; Pfeffer & Veiga, 1999). It is simply because high performing organizations have the resources to divert to the development activities that low performing find difficult to tend to. Using this logic, we can expect profitable organizations in Pakistan to invest more in developing effective HRM systems thereby exposing employees to new alternatives in HRM. Thereby, employees in these organizations are expected to receive extensive training and have better developed skills. This makes these employees more marketable than their counterparts in lower- performing organizations. In a competitive environment, these employees from higher performing organizations, with a greater awareness, higher expectation of development- oriented HRM systems, and better developed skills, are more likely to hop jobs if they are not satisfied with HR practices in their respective organizations. Thus we hypothesize:

Hypothesis 5. Turnover intentions of employees in high performing organizations are more

likely to be negatively influenced by increased HRS.

With the help of these hypotheses we hope to develop a better understanding of SHRM issues in Pakistan. In addition, highlighting demographic and organizational antecedents of turnover, testing a relationship between HRS and turnover as well as an investigation of differences between multinationals and local companies will help us capture a snap shot of cultural transformations in a developing country like Pakistan.

3. Method

Having laid out a theoretical premise of the expected relationships among turnover intention, HRS, and other individual and organizational variables, we will now discuss the methodology that was employed to test these hypotheses.

3.1. Sample and data collection

Becker and Gerhart (1996) argue that conducting research in a diverse sample of organizations having different objectives may in some cases cause a researcher to deviate, since some organizations may focus upon profit while others may strive for market share. On the other hand, single-industry studies enable researchers to assess the unique strategies (both business and HR) that organizations engage in as a means of improving performance. Since one of our key hypotheses involves investigating the impact of high performance we decided to utilize a single industry analysis. This decision also allowed us the opportunity to keep the industrial structure, and environment factors (according to Glebbeek & Bax’s, 2004 categories of moderating variables related to turnover) constant. We selected the banking industry because it is highly competitive and operates in a dynamic environment of growth (through innovation of products, technology and

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395384

services). We think for the purpose of an HRS and turnover intention study, this industry was useful because it represented an employee’s market and high workforce mobility, constituting a fertile ground for managers looking to improve levels of HRS in order to retain employees. The total population in the Pakistani financial industry was 17 organizations. Of these

12 decided to participate in the current study, thus the current sample represents 70% of the entire population. The sample included six multinationals and six locally owned companies. The first author contacted all 17 CEO in the industry via a formal letter and/or a telephone call and briefly described the objective of the study. She then requested an appointment to discuss her research further. At this meeting she explained the entire research process and extended an invitation for the organization to participate in the study. Twelve of these CEOs accepted the offer, and referred her to the HR manager or other area managers. She used office locations of second and third contact managers in four main cities of Pakistan (Karachi, Lahore, Rawalpindi and Islamabad) to personally distribute 75–100 questionnaires. She tried to establish direct contact with many employees by walking around and talking with them. She explained the objective of the research to each respondent in order to establish trust, to be regarded positively and to appeal to them on the basis of the usefulness of the study, as has been explained by Dilman (1978). This strategy also allowed her the flexibility to select a sample according to the primary criteria; a large representation of non-managers and employees from a wide cross section of offices, location and functions. Free mingling with the employees also helped her collect content- specific information relating to various HR practices. This will be reported later in the paper. Respondents had the choice to hand back the questionnaire to her during the same visit, the next visit (which took place after one week) or to mail it back in a self addressed envelope. A majority of the questionnaires were personally returned to her. Since questionnaires were distributed personally, it was considered feasible to reach

75–100 employees in each organization, with an intentional split of 30% managers and 70% non-managers in order to comply with the main objective of collecting data from non-managerial staff. Some organizations and the offices that the first author was allowed access to were smaller than others. Hence a total of 977 questionnaires were distributed. Of these 593 were returned, and only 508 were usable (representing an average of 42% of respondents and a range of 40–66 respondents per organization) for an effective response rate of 51%. All of these were Pakistani employees in keeping with the objective of present study. Among the 508 responses, 69% of the respondents were non-managers, and 60% of them were men. The average age of this sample was 37 and education was 15 years of schooling (which translates into at least a college degree in Pakistan). Their tenure in their current organization was two to five years, thus the likelihood of respondents experiencing all aspects HRM in their organizations was quite high.

4. Measures

As explained previously, current research focuses upon using six variables in order to analyze the relationship between turnover intention and HRM. These include one independent variable, i.e. HRS, one dependent variable, which is turnover intention, and four moderators, including employee age, gender, organizational type including performance and ownership. The following discussion describes specific items and measures used for each of these variables.

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395 385

4.1. HR satisfaction (HRS)

The scale for HRS was based on Delery and Doty’s work (1996). It had a total of 10 item practices, each measured with a five-point Likert scale (from strongly satisfied to strongly dissatisfied), a higher score reflecting a higher level of HRS. These items included training for skill development, recruitment procedures, opportunities for higher learning, opportunities for promotion, compensation, performance appraisal system, job security, management support, relationship with the manager, and recognition for a good job. The Crobanch coefficient alpha of this scale is .90, which makes the scale reliable to use.

We have previously argued that HRS, as it is conceptualized in this paper, is different from job satisfaction. However, in order to make sure HRS was a distinct measure, we decided to statistically test its validity and included job satisfaction in our analysis. We used three items, ranging from overall satisfaction on job, job demands, and job significance for this measure. For each item in this scale, respondents indicated how satisfied they were with their present job in a 5-point response format ranging from very dissatisfied to very satisfied. The internal consistency and reliability of measure of job satisfaction was found to be .62, and the correlation coefficient between this scale and HRS scale was calculated as .64 ðpo:001Þ. This is modest and demonstrates that HRS and job satisfaction were two distinct and independent variables. Furthermore, we performed a paired-sample t test to analyze the difference between job satisfaction and HRS. Again, the result showed these two variables to be significantly different (the mean difference was d ¼ :09, t ¼ 2:97, po:01), leading us to argue that HRS is not the same as job satisfaction.

4.2. Turnover intention

In an effort to understand the ways in which employees were responding to a rapidly growing economic environment and changing HRM systems, we decided to study turnover intentions as a dependent variable. This, as explained previously, fitted well with the motivation of present study and helped us investigate employee experiences. We also found several of the sample companies had no formal record of voluntary employee turnover, thus, it inadvertently turned out to be a good measure for the study. We used one item, ‘‘Sometimes I feel like leaving this organization’’ to measure this variable on a 5-point Likert scale ranging from 1 as ‘‘strongly disagree’’ to 5 as ‘‘strong agree’’. A higher score reflected a stronger turnover intention. Since turnover intention is a clearly defined by the question itself, a single-item measure was considered sufficient to measure it (Sackett & Larson, 1990; Wanous, Reichers, & Hudy, 1997).

4.3. Organizational performance

In order to obtain a reliable performance variable, three indicators of organizational performance were used: a percentage increase of profits, a percentage increase of market share and a percentage increase of deposit base. We chose to measure profits because it is the most commonly understood financial measure in the selected industry. The latter two accounted for the common growth strategy that each bank was pursuing at time of data collection, and reflected our sample’s directed efforts and commitment to achieving this strategy. Our choice of market-based measures was also guided by the existing argument emphasizing their superiority, and because human based capital competencies are in part the

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395386

source of ‘intangible assets’ represented by the market value of the firm (Becker, Huselid, Pickus, & Spratt, 1997; Guthrie & Datta, 2004). A composite scale, rather than a single indicator of profit, was used for two reasons. First,

it was better able to capture the comprehensive market performance of the sample in terms of its strategy, which was guiding organizations’ HR efforts. Second, the use of a composite scale for organizational performance is quite common in SHRM, and is expected to enhance comparability of our findings (Becker and Gerhart, 1996; Huselid, 1995). The reason for measuring the percentage increase for the above indicators, rather than the

total amount, was two fold. First it allowed us to control for organizational size since smaller banks (no matter how successful) were likely to have a smaller deposit base and profits. Second, such a strategy helped us focus upon the sample’s progress over a one-year period, and allowed us the opportunity to study changes in performance and turnover intentions guided by HRS. Several other scholars in the field have pursued a similar strategy of using multiple year means, ratios and percentages (Chang & Singh, 1999; Hambrick & Abrahamson, 1995; Rajagopalan & Datta, 1996). Annual reports and information from newspapers and magazines (in cases annual reports were not available) were used to obtain data for chosen performance indicators. This information is considered to be accurate and reliable as these articles referenced company published reports and data. Since these three indicators were on different scales (local currency amount or percentage), standardized scores of these three indicators were combined to create one variable—organizational performance. The reliability of this composed variable, in terms of Crobanch’s alpha, was found to be .88, and reflected that it was an appropriate measure of analysis. The next step was to categorize organizations into successful and non-successful

companies in order to test the above hypotheses. We categorized low performing organizations (standardized scores were lower than 0) as 1 and high performing organizations 2 (standardized scores were higher than 0).

4.4. Age

Age was an important moderator because one of the primary objectives of present analysis was to determine the effect of age on HRS and turnover intention relationship. We used three categories: 1 represents ages below 30 years; 2 represents the age group of 31 to 42 years old; and 3 representing ages 43 and older. This age division was based upon previous research indicating that the opening up of the business environment within Pakistan since 1990 resulted in a mushroom growth of business schools within the country and led to an introduction of American management education (Khilji, 2004). We believe since employees younger than 42 years of age would have been exposed to management education earlier on in their professional lives, they would be more likely to have adopted modern work related values. This group was further split into two categories, with the expectations that the below 30 years old group is more readily exposed to modern values in their teens. The ratio of these three groups of employees for the sample was 57:27:16 for younger, middle and older groups respectively.

4.5. Control variables

The statistical analysis was conducted at the individual level. In order to establish the robustness of the hypothesized relationships, we controlled organizational size in all

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395 387

analyses. Evidence regarding size is contradictory. Zhu (2000) suggests that larger organizations in terms of total number of employees are likely to possess and invest far greater resources for the development of effective HR systems that will attract and retain a capable workforce. This factor may positively influence turnover intention of employees. However, Khilji (2002) in her discussion of Pakistani organizations has argued commitment of top managers is a far more important determinant; and unless an organization’s managers are committed to changes, positive employee attitudes cannot be gained. Unsure of its impact on our sample, we decided to control the effect of organizational size.

5. Results

We conducted hierarchical regression to test the direct impact of HRS on turnover intention and used serials of ANCOVA to test the moderating effects of the aforementioned measures. Hierarchical regression was deemed most useful, because it specifies the order in which variables are entered and removes the spurious relationship before checking for the contribution of independent variables against the explanation of the dependent variables (Cohen & Cohen, 1983). It also helped us rule out the effect of organizational size which, as explained previously, served as a control variable.

Table 1 presents the descriptive statistics, reliability coefficients and correlations for all variables. The values presented in this table are at the aggregate level with the total sample. All reliability coefficients (as) are acceptable. As expected, the independent variable, HRS, negatively correlated with the dependent variable, turnover intention. Other correlation coefficients were also found to be within the expected ranges and directions.

Hypothesis 1 predicted that HR satisfaction will negatively influence the turnover intention. In order to test this hypothesis, hierarchical regression analysis was conducted (Cohen & Cohen, 1983) by entering the control variable (organization size) at the first step and entering the independent variable (HR satisfaction) at the second. The increased effect of HR satisfaction on the turnover intention was found significant (standardized b ¼�:45, DR2 ¼ :21, po:001). Therefore, Hypothesis 1 was confirmed.

Table 1

Descriptive data, reliability and correlation coefficients

M Sd. a 1 2 3 4 5 6

1. Organization

size

2372 4657 n/a

2. Age 1.59 .75 n/a .24 ***

3. Gender 1.17 .38 n/a �.08 �.19 ***

4. Organization

type

1.81 .40 n/a .36 ***

.10 *

�.06

5. Organization

performance

1.41 .49 .88 �.31 ***

�.11 **

�.03 .01

6. HR satisfaction 3.35 .61 .90 �.01 .21 ***

�.09 *

.04 .17 ***

7. Turnover

intention

3.01 1.25 n/a �.02 �.18 ***

.08 �.04 �.10 *

�.45 ***

N ¼ 523.

Note: *po:05; **po:01; ***po:001.

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395388

Hypotheses 2–5 predicted the moderating effects of age, gender, organizational type, and organizational performance on the relationship between HRS and turnover intention. In order to test these hypotheses, a set of univariate covariance analyses (ANCOVA) were conducted controlling for organization size. Hypothesis 2 predicted that age would be a moderator for the HRS and turnover

intention relationship, in the sense that younger employees would have a stronger HRS and turnover intention relationship. The ANCOVA analysis, after controlling the main effects of organization size, the independent variable and the moderator, did show that age has interaction effect with HRS on turnover intention (F ¼ 5:03, o2 ¼ :02, po:01). In order to understand this result better, mapping of the interacting effect between age and HR satisfaction was done (Refer to Fig. 1). Clearly Hypothesis 2 was confirmed. The slopes for younger employees (for both groups under age 42 years old) are much steeper than that for older employees (over 43 years old). The difference between younger groups (under 31 years old group and between 32 and 42 years old group) was not found to be significant and a follow-up test confirmed this: F ð424;1Þ¼ :15, o2 ¼ :00, p4:05. Hypothesis 3 predicted the moderating effect of gender on the HRS and turnover

intention relationship that the male employees would be more affected by HRS. However, ANCOVA test did not confirm this hypothesis (see Table 1, F ¼ 2:24, o2 ¼ :00, p4:05). Similar results were found for Hypothesis 4, which predicted that multinational company employees would have a stronger HRS and turnover intention relationship. Thus Hypothesis 4 was also rejected (F ¼ :59, o2 ¼ :00, p4:05). Hypothesis 5 predicted that the HRS and turnover intention relationship will be

stronger for employees working for higher performing organizations. Results confirmed this hypothesis (F ¼ 5:69, o2 ¼ :01, po:05). Fig. 2 further illustrated the moderating effect of organization performance on HRS and turnover intention relationship and depicted a steeper slope of HRS and turnover intention correlation for higher performing organizations.

1

2

3

4

5

1 3

HR Satisfaction

T u

r n

o v

e r I

n te

n ti

o n

Below 30 year Old

31-42 Years Old

Over 43 Years Old

4 52

Fig. 1. Age as a moderator for HRS and turnover relationship.

ARTICLE IN PRESS

1

2

3

4

5

1 2 3 4 5

HR Satisfaction

T u

r n

o v e r I

n te

n ti

o n

High on Organization

Performance

Low on Organization

Performance

Fig. 2. Organization performance as a moderator for HRS and turnover relationship.

S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395 389

6. Discussion

We conducted present research in order to evaluate the critical value of HRS in SHRM debate and to highlight the significance of a multiple-respondent strategy. We tested the impact of HRS on turnover intention, and explored the moderating effects of several non- job related (i.e. individual and organizational) variables in a vibrant economy such as Pakistan. Statistical analyses reveal interesting results. Findings support our primary hypothesis that higher HRS levels reflect lower turnover intentions even in an emerging country like Pakistan. Of the four variables we chose as moderators, age and organizational performance emerged as significant moderators of the HRS-turnover relationship, implying younger employees’ decision to turnover is more likely to be influenced by levels of HRS. We also found that turnover intention of employees in high performing organizations is more likely to be determined by their respective levels of HRS. Together these findings underscore the importance of maintaining higher levels of HRS within organizations.

The moderating effects of gender and organization type on the HRS and turnover intention relationship were not confirmed. In order to interpret this result, we referred back to the dataset and discovered it included a small sample of females, a majority of whom were young. Among a total of 86 female respondents (out of 508), five were in the ‘‘over 43 years old’’ category, 14 in the ‘‘32–42 years old category’’ and 67 in the ‘‘under 30 years old category’’. It is possible for such small numbers to contaminate findings. At the same time, we also believe our initial expectation, upon which we based our hypothesis, that the women are less likely to be influenced by HR practices in their organizations

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395390

because of family pressures may be flawed for a newer generation of women. The young female employees that the first author met during the process of conducting surveys were enthusiastic, motivated and committed to their professions and careers. In informal conversations they discussed they had worked hard to obtain MBAs and additional banking accreditations to stay competitive in the industry. Married or not, they said their families were supportive of their careers and professional ambitions. A respectable percentage of them had already attained managerial positions. This contradicts our initial impression of Pakistan where we expect women to primarily fulfill domestic roles. With respect to a failure at finding the moderating effect of the organization type, we

offer a range of possible interconnected explanations, based upon institutional theory, ‘emic’ aspects of culture, high quality/education of local workforce and the level of management commitment, as an internal resource base (Kamoche, 1997). We will begin with providing historical explanations, and then describe contemporary changes precipitated by market reforms of the 1990s to present a picture of the current conditions. First, all of the chosen multinational companies had been operating in Pakistan for more than 40 years. There were very few expatriates, and none in top leadership positions. This is likely to lead to the dominance of certain cognitive sets of HR knowledge across both multinationals and local companies (Gomez & Sanchez, 2005) and expected to limit transfer of practices from parent company to the subsidiary to result in a dominance of local characteristics (Tsang, 1999). Furthermore, strict regulatory environment during nationalization and periodic stagnation of the 1970s and the 1980s influenced multi- national to conform their HR practices to those of the local organizations in an effort to gain institutional legitimacy (Rosenzweig & Nohria, 1994; Tanure & Duarte, 2005). However, market reforms of the 1990s invigorated HRM systems across both multi- nationals and local organizations. A new generation of employees, influenced by the market culture and reforms, in addition to direct exposure to electronic mass media, and foreign investment, emerged. Since ‘emic’ aspects in an organization are based upon employees’ work-related values, and HRM is closely anchored in assumptions held by senior managers (Tayeb, 1994), this generation of employees played a vital role in pushing organizations towards developing new HRM structures and policies across organizations (Khilji, 2004). Our sample was no exception. Conversations with several employees, during the process of questionnaire distribution and collection, revealed each organization, irrespective of its nationality and origin, had adopted MBO, performance-based remuneration, merit-led recruitment, and extensive training. It can thus be argued that recent contextual changes, including transitions in employees’ work-related values and rapid economic development in Pakistan, have mitigated the institutional effects of the pre-reform era, and created an environment where both multinationals and local organizations are adopting SHRM systems to achieve and maintain a competitive advantage. It can also be argued that since Pakistan possesses an educated workforce (although the moderating effect of education on turnover and HRS, which we calculated later, was found to be insignificant and we attribute this to a highly educated sample as explained in Section 3) with adequate managerial potential, opportunities relating to innovation in HRM are equally available to multinationals and local organizations. For example, a recent study by Khilji (2002) reveals that some local companies (rather than multinationals), managed by foreign qualified and experienced HR managers and CEOs were the first to take the initiative to develop new HR programs. Management in these local organizations was also found to be far more committed to transforming the

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395 391

traditional corporate culture in order to achieve competitive advantages. Present study therefore supports existing evidence relating to the significance of organizational culture in the development of innovative products, and potential (Lau & Ngo, 2004; Solvay & Sanglier, 1998). Additionally, it locates the internal resource environment at the heart of SHRM (Kamoche, 1997). Since all organizations included in our sample had adopted innovative HRM, we can argue that the previously discussed distinction between multinationals and local companies is not likely to emerge as a significant moderator. Finally, confirmation of the hypothesis that high performing organizations have a stronger HRS and turnover relationship reaffirms our explanation that it is not the organizational type (multinational or local), but the management commitment and effective utilization of resources that helps determine the relationship between HRS and turnover intentions in Pakistan (Keenoy, 1999; Khilji, 2002; Storey, 1992).

The above discussion is not intended to overplay the significance of our findings. In fact it is presented to highlight the complexity of HRS and turnover decisions in organizations and indicates the interplay of several contextual variables, relating to labor conditions, organizational culture, and even history of business environment and regulations, in explaining the relationship between of HRS and turnover. Thus it is apparent that, far and beyond age and organizational performance, both external factors and the internal resource environment are also important determinants leading to high HRS and low turnover. This explanation supports the contextual paradigm in SHRM (Brewster, 1999). It is worth clarifying that while an understanding of modern HRM prevailing in Pakistan has been influenced by western thought and practice, and the basic concept of SHRM, in view of our findings relating to a significant relationship between HRS and turnover and moderating effects of high performance organizations and age on HRS and turnover, appears to have much in common with several studies conducted in the west (these have been cited throughout this paper), implementation of HR practices is distinctly local, and shaped by external as well as internal organizational environment. This is an important finding, and supports the notion of ‘‘crossvergence’’ in HRM ((McGaughey and De Cieri, 1999; Khilji, 2002), which recognizes the competing influences of both global and local factors in international management. The present analysis also emphasizes the need to broaden the focus of turnover research beyond attitudinal, demographic and relational factors. Hopefully it will encourage future research to explore contextual issues in answering the turnover question which has challenged social scientist and practitioner for years, ‘‘why do people leave?’’.

7. Conclusions and limitations

Despite various important findings mentioned above, we admit that this study has several limitations. First, the sample size at the organizational level is small, even though we have a good sample size at the individual level. We admit this aspect hindered our ability to conduct a cross-level analysis and comparison of sample organizations. Second, although a single industry analysis has its merits its generalizability can be questionable. A multi-industry study is better able to facilitate the understanding of issues in a broader setting. Third, a single-country analysis limits our ability to generalize findings to a relationship between HRS and turnover in other international settings. Fourth, both turnover and organizational performance measures were cross-sectional measures, and limited our ability to discuss cause and effect of the turnover and HRS relationship.

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395392

Finally although we emphasized the present study aimed at generating a fuller understanding of turnover, only turnover intentions were used in the analysis. Additionally, these were measured with a one-item scale. We are aware of the limitation in capturing individuals’ complete intent to leave as a result of HR dissatisfaction. It is, therefore, imperative that future studies focus on multiple industries, larger samples and multiple-item constructs in analyses to overcome these weaknesses. The present analysis has important implications for studies devoted to understanding

SHRM in developing countries. It demonstrates that the basic concept of SHRM, i.e. effective HR practices can lead to higher organizational performance, is universal and just as applicable in a developing country. However at the same time, explanations of several findings above signify relevance of contextual factors within organization and its environment. The study also supports previous research, which showed younger employees in India, China, Pakistan and Central America have expectations of egalitarian and involvement-oriented HRM, even though these run counter to their cultural norms (Chatterjee & Pearson, 2000; Khilji, 2002; Osland & Osland, 2005). Both a study of HRS and a multiple respondent research strategy have therefore been useful in capturing the true nature of implementation of HR practices. Making the argument that the day-to-day experiences of employees are critical has served the existing SHRM debate well, and has contributed to a better understanding of issues relevant to international SHRM debate. By highlighting the significance of several contextual factors, this study also hopes to expand the focus of turnover research.

References

Arthur, J. B. (1994). Effects of HR systems on manufacturing performance and turnover. Academy of

Management Journal, 37(2), 670–687.

Baruch, Y., & Budhwar, P. (2006). A comparative study for career practices for management staff in India and

Britain. International Business Review, 15(1), 84–101.

Becker, B. E., & Gerhart, B. (1996). The impact of HRM on organizational performance: Progress and prospects.

Academy of Management Journal, 39(4), 779–801.

Becker, B. E., Huselid, M., Pickus, P., & Spratt, M. (1997). HR as a source of shareholder value: Research and

recommendations. Human Resource Management, 36(1), 39–47.

Brewster, C. (1999). Strategic human resource management: The value of different paradigms. Management

International Review, 39(3), 45–64.

Campbell, D. J., & Campbell, K. M. (2003). Global versus facet predictors of intention to quit: Differences in a

sample of male and female Singaporean managers and non managers. The International Journal of Human

Resource Management, 14(7), 1152–1177.

Chang, S. J., & Singh, H. (1999). The impact of modes of entry and resource fit on modes of exist by multi-

business firms. Strategic Management Journal, 20(11), 1019–1035.

Chatterjee, S., & Pearson, C. (2000). Indian managers in transition: Orientations, work goals, values and ethics.

Management International Review, 40(10), 81–95.

Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.).

NJ: Lawrence Erlbaum Associates.

Cooke, F. L. (2005). HRM, work and employment in China. New York, NY: Routledge.

Cotton, J. L., & Tuttle, J. F. (1986). Employee turnover: A meta-analysis and review with implications for

research. Academy of Management Review, 11(1), 55–70.

Datta, D., Guthrie, J., & Wright, P. (2005). Human resource management and labor productivity: Does industry

matter? Academy of Management Journal, 48(1), 135–145.

Delery, J. E., & Doty, D. H. (1996). Modes of theorizing in strategic HRM: Tests of universalistic, contingency,

and configurational performance predictions. Academy of Management Journal, 39(4), 802–835.

Dilman, D. A. (1978). Mails and telephone surveys: The total design methods. New York: Wiley.

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395 393

Duarte, R. G. (2001). Cross-border acquisitions and change in domestic management practices—The case of Brazil.

Ph.D. dissertation, Judge School of Business, University of Cambridge, UK.

Economist. (2006). Emerging market indicators. The Economist, 4(March 2006), 98.

Elaine, M. (1997). Job tenure shifts for men and women. HR Magazine, 42(5), 20.

Fang, B. (2006). Super shoppers: How yuppies consumers in China and India and transforming the world

economy. The US News, May 1, 2006.

Glebbeek, A., & Bax, A. H. (2004). Is employee turnover really harmful? An empirical test using company

records. Academy of Management Journal, 47(2), 277–286.

Gomez, C., & Sanchez, J. I. (2005). HR’s strategic role within MNCs: Helping build social capital in Latin

America. The International Journal of Human Resource Management, 16(2), 2189–2207.

Government of Pakistan. (2005). Economic Survey of Pakistan 2004– 2005. Retrieved May 3, 2006, from /http:// www.finance.gov.pk/survey/home.htmS.

Gratton, L., & Truss, C. (2003). The three-dimensional people strategy: Putting human resources policies into

action. Academy of Management Executive, 17(3), 74–86.

Guthrie, J. P. & Datta, D. K. (2004). HRM and firm performance: Peeling back the onion. Paper presented at

Academy of Management meeting, New Orleans, LA.

Hambrick, D. C., & Abrahamson, C. (1995). Assessing managerial discretion across industries: A multi-method

approach. Academy of Management Journal, 38(5), 1427–1441.

Harrison, D. A., Newman, D. A., & Roth, P. A. (2006). How important are job attitudes? Meta analytic

comparisons of integrative behavioral outcomes and time sequences. Academy of Management Journal, 49(2),

305–326.

Herbert, J. D., Chatterjee, S., & Heuer, M. (2006). Management in India: Trends and transition. New Delhi: Sage

Publications.

Hom, P. W., Caranikas-Walker, F., Prussia, G. E., & Griffeth, R. W. (1992). A meta-analytical

structural equation of analysis of model of employee turnover. Journal of Applied Psychology, 77(6),

890–909.

Hom, P. W., & Griffeth, R. W. (1995). Employee turnover. Ohio: Southwestern College Press.

Hom, P. W., & Kinicki, A. D. (2001). Toward a greater understanding of the how dissatisfaction drives employee

turnover. Academy of Management Journal, 44(5), 975–987.

Huselid, M. A. (1995). The impact of HRM practices on turnover, productivity, and corporate performance.

Academy of Management Journal, 38(3), 635–672.

Iverson, R., & Pullman, J. (2000). Determinants of voluntary turnover and layoffs in an environment

of repeated downsizing following a merger: An event history analysis. Journal of Management, 26(5),

977–1003.

Kamoche, K. (1997). Research Note: Managing human resources in Africa: Strategic, organizational and

epistemological issues. International Business Review, 6(5), 537–558.

Keenoy, T. (1999). HRM as hologram: A polemic. Journal of Management Studies, 36(1), 1–23.

Khatri, N., Budhwar, P., & Fern, C. T. (2005). Employee turnover: Bad attitude or poor management? Retrieved

September 3, 2005, from /http://www.ntu.edu.orgS. Khilji, S. E. (2002). Modes of convergence and divergence: An integrative view of multinationals in Pakistan. The

International Journal of Human Resource Management, 13(2), 232–253.

Khilji, S. E. (2004). Wither tradition: An evidence of generational differences in human resource satisfaction from

Pakistan. The International Journal of Cross Cultural Management, 4(2), 141–156.

Khilji, S. E., & Wang, X. (2006). Intended and implemented HRM: The missing lynchpin in

strategic human resource management. The International Journal of Human Resource Management, 17(7),

1171–1189.

Lau, C. M., & Ngo, H. Y. (2004). The HR system, organizational culture and product innovation. International

Business Review, 13(6), 685–703.

Lawler, E. (1986). High-involvement management. SF: Jossey Bass.

Lawler, E. (2005). Creating high-performance organizations. Asia Pacific Journal of Human Resource

Management, 43(1), 10–17.

Lee, T. W., & Mitchell, T. R. (1994). An alternative approach: The unfolding model of voluntary employee

turnover. Academy of Management Review, 19(1), 51–89.

Lee, T. W., Mitchell, T. R., Holtom, B. C., McDaniel, L., & Hill, J. W. (1999). The unfolding model of voluntary

turnover: A replication and extension. Academy of Management Journal, 42(4), 450–463.

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395394

Maertz, C. P. (2001). Why employees stay or quit an organization. Paper presented at the 61st annual meeting of

the Academy of Management, Washington, DC.

March, J. G., & Simon, H. A. (1958). Organizations. New York: Wiley.

McGaughey, S., & De Cieri, H. (1999). Reassessment of convergence and divergence dynamics: Implications for

international human resource management. International Journal of Human Resource Management, 10(2),

235–250.

Mitchell, A. (2005). Revival of startups: The new trend. E-commerce Times. Retrieved May 17, 2005, from

/http://www.EcommerceTimes.comS. Mitchell, T. R., Holton, B. C., Lee, T. L., Sablynski, C. J., & Erez, M. (2001). Why people stay:

Using job embeddedness to predict voluntary turnover. Academy of Management Journal, 44(6),

1102–1121.

Osland, A., & Osland, J. S. (2005). Contextualization and strategic international human resource management

approaches: The case of Central America and Panama. The International Journal of Human Resource

Management, 16(12), 2218.

Pfeffer, J. (2005). The myth of disposable worker. Business 2.0. October 2005.

Pfeffer, J., & Veiga, J. F. (1999). Putting people first for organizational success. Academy of Management

Executive, 13(2), 37–48.

Phillips, J. (1996). Accountability in Human Resource Management. Houston: Gulf Publishing.

Purcell, J. (1999). Best practice and best fit: Chimera or cul-de-sac? Human Resource Management Journal, 9(3),

26–41.

Rajagopalan, N., & Datta, D. K. (1996). CEO characteristics: Does industry matter? Academy of Management

Journal, 39(1), 197–215.

Rosenzweig, P. M., & Nohria, N. (1994). Influences on human resource management in multinational

corporations. Journal of International Business Studies, 25(2), 229–252.

Sackett, P. R. & Larson, J. R. (1990). Research strategies and tactics in organizational and industrial psychology.

In M.D. Dunnette, & L.M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 1, 2nd

ed., pp. 419–489). Palo Alto, CA: Consulting Psychologists Press.

Shaw, J. D., Delery, J. E., Jenkins, G. D., Jr., & Gupta, N. (1998). An organizational-level analysis of voluntary

and involuntary turnover. Academy of Management Journal, 41(5), 511–525.

Solvay, J., & Sanglier, M. (1998). A model of the growth of corporate productivity. International Business Review,

7(5), 463–481.

Storey, J. (1992). Developments in the management of human resources. Oxford: Blackwell.

Tanure, B., & Duarte, R. G. (2005). Leveraging competitiveness upon national culture traits: The mana-

gement of people in Brazilian companies. The International Journal of human resource management, 16(12),

2201–2217.

Tayeb, M. (1994). Organization and cultures: Methodology reconsidered. Organization Studies, 15(3),

429–446.

Taylor, S., Beechler, S., & Napier, N. (1996). Towards an integrative model of strategic international human

resource management. Academy of Management Review, 21(4), 959–985.

Truss, C. (2001). Complexities and controversies in linking HRM with organizational outcomes. Journal of

Management Studies, 38(8), 1121–1149.

Tsang, E. (1999). The knowledge transfer and learning aspects of international HRM: An empirical study of

Singapore MNCs. International Business Review, 8(5–6), 591–609.

UNCTAD (United Nations Conference on Trade and Investment). (2005). World Investment Report. Retrieved

March 2, 2006 from /http://www.unctad.org/Templates?WebFlyer.asp?intItemID=3489&lang=1S. Wai, C. T., & Robinson, C. D. (1998). Reducing staff turnover: A case study of dialysis facilities. Healthcare

Management Review, 23(4), 21–42.

Walton, R. E. (1985). Towards a strategy of eliciting employee commitment based on policies of mutuality. In R.

E. Walton, & P. R. Lawrence (Eds.), HRM: Trends and challenges (pp. 35–65). Boston: Harvard University

Press.

Wanous, J. P., Reichers, A. E., & Hudy, M. J. (1997). Overall job satisfaction: How good are single-item

measures? Journal of Applied Psychology, 82(2), 247–252.

Warner, M. (2004). Human resource management in China revisited: Introduction. The International Journal of

Human Resource Management, 15(4), 617–634.

Weisberg, J., & Kirschenbaum, A. (1993). A model of voluntary turnover among hospital CEOS. Hospital and

Health Service Administrative, 40(3), 362–385.

ARTICLE IN PRESS S.E. Khilji, X. Wang / International Business Review 16 (2007) 377–395 395

Wright, P. M., & McMahan, G. C. (1992). Theoretical perspectives for strategic human resource management.

Journal of Management, 18(2), 295–320.

Wright, P., Scott, A., Snell, A., & Dyer, L. (2005). New models of strategic HRM in a global context. International

Journal of Human Resource Management, 16(6), 875–881.

Wright, P., & Snell, S. A. (1991). Towards an integrative framework of strategic human resource management.

Human Resource Management Review, 1(3), 203–225.

Zhu, Y. (2000). HRM in Taiwan: An empirical case study. Human Resource Management Journal, 10(4),

32–45.

Zhu, Y., & Warner, M. (2000). An emerging model of employment relations in China: A divergent path from the

Japanese? International Business Review, 9(3), 345–361.