ꇲ 냪ꗟ꒤ꑳꑪ뻇 ꑈꑏ룪랽뫞뉺 곣ꡳ꧒ Institute of Human Resource Management National Sun Yat-sen University 돕ꑨ뷗ꓥ Doctoral Dissertation ꑈꑏ룪랽룪ꅂꑈꑏ룪ꖻ뭐닕슴셚껄꒧곣ꡳ A Study of Human Resource Investment, Human Capital, and Firm Performance 곣ꡳꗍꅇ뢭탔뫕 벶 Chu-Chen Rosa Yeh 뻉뇐뇂ꅇ뎯ꕀ귵 Shyh-Jer Chen ꒤뗘ꗁ냪‹㔠꙾‷ꓫ July 2006?
Table of Contents Table of Contents........................................................................................................................i List of Tables.............................................................................................................................iii List of Figures...........................................................................................................................iv Acknowledgement.....................................................................................................................v Abstract.....................................................................................................................................vi Chapter I: Introduction...............................................................................................................1 Rationales of the Study......................................................................................................2 Research Questions............................................................................................................5 Research Objectives...........................................................................................................6 Chapter II: Literature Review....................................................................................................8 Resource-Based View of the Firm.....................................................................................8 Strategic Human Resource Management.........................................................................11 Human Capital as a Black Box between HRM and Firm Performance...........................18 Summary..........................................................................................................................40 Chapter III: Theoretical Framework & Hypotheses................................................................41 Universalistic Model of HR ROI.....................................................................................42 Configurational Model of HR ROI..................................................................................46 Summary..........................................................................................................................52 Chapter IV: Methodology........................................................................................................54 Research Design...............................................................................................................54 Variable Definition and Measurement.............................................................................62 Scale Validation................................................................................................................67 Summary..........................................................................................................................83 Chapter V: Analysis and Results..............................................................................................84 Correlations......................................................................................................................84 Hierarchical Regression...................................................................................................86 Summary..........................................................................................................................94 i
Chapter VI: Discussions and Conclusions...............................................................................96 Discussions.......................................................................................................................96 Implications....................................................................................................................100 Contributions..................................................................................................................104 Limitations.....................................................................................................................105 Future Research Suggestions.........................................................................................107 Conclusions....................................................................................................................109 References..............................................................................................................................111 Appendix I..............................................................................................................................121 Appendix II............................................................................................................................126 Appendix III...........................................................................................................................127 ii
List of Tables Table 1. Definitions of human capital................................................................................22 Table 2. Classification of human capital constructs..........................................................25 Table 3. Popular human capital measurements..................................................................30 Table 4. Human capital constructs: organizational level vs. individual level....................36 Table 5. Research hypotheses............................................................................................53 Table 6. Sample profile......................................................................................................58 Table 7. HR investment measurement model parameter estimates...................................70 Table 8. Goodness of fit statistics of final HR investment measurement models.............71 Table 9. Human capital measurement model parameter estimates....................................74 Table 10. Goodness of fit statistics of final human capital measurement models...............75 Table 11. Firm performance measurement model parameter estimates..............................79 Table 12. Goodness of fit statistics of final firm performance measurement models.........80 Table 13. Construct reliability..............................................................................................83 Table 14. Means, standard deviations, reliabilities, and correlations..................................85 Table 15. Results of regression analyses of universalistic model hypotheses.....................86 Table 16. Regression results of human capital dimensions on HR investment portfolios...88 Table 17. Regression results of operational performance on dimensions of HC.................90 Table 18. Regression results of financial performance on dimensions of HC.....................90 Table 19. Regression results of market performance on dimensions of HC........................91 Table 20. Collinearity statistics of regression models.........................................................94 Table 21. Results of hypothesis testing................................................................................95 i ii
List of Figures Figure 1. The resource-based view of the firm....................................................................8 Figure 2. Research framework and hypotheses.................................................................41 Figure 3. Final two-factor HR investment measurement model.......................................69 Figure 4. One-factor HR investment measurement model................................................72 Figure 5. Final four-factor human capital measurement model........................................73 Figure 6. Three-factor human capital measurement models.............................................76 Figure 7. Two-factor human capital measurement models...............................................77 Figure 8. One-factor human capital measurement model.................................................78 Figure 9. Final three-factor firm performance measurement model.................................79 Figure 10. Two-factor firm performance measurement models..........................................81 Figure 11. One-factor firm performance measurement model............................................82 iv
Acknowledgement Writing a dissertation has proved to be a truly humbling process for me, a critically needed shock therapy when I was becoming too full of myself after years of industry and consulting experience. I now understand how difficult it is to do even just an adequate research with the amount of intelligence, effort, support, perseverance and sometimes a little bit of luck that must go behind it. I have a new found respect to all who thrive in the field of academic research after this learning experience. Gratitude is extended to those who helped during this period of time, the faculty members whom I took courses from, my classmates whose friendship and camaraderie I’ll always cherish, and my junior peers who are always ready to lend a hand. I am especially grateful to my advisor, Dr. Shyh-Jer Chen, who gave me the freedom to set my own course and yet kept me grounded on the basics of research, and to my committee members who had to bear with my inexperienced academic research skills and still provided valuable inputs to this research. I dedicate this dissertation to my dear husband, Dr. Yu-Hui Tao, who has been the constant driving force in my adult life, always pushing (well, he would prefer the word “encouraging”) me to strive for the impossibles to better myself. He supported me through my pursuit of the doctoral degree not only by sharing the house chores, taking care of the little ones, and providing food on the table, but also by setting a perfect role model in his positive outlook of the life, in maintaining his integrity as a teacher, in his relentless pursuit of efficiency and order, and in his love and affection toward his families. Without him I could never have gotten this far. Chu-Chen Rosa Yeh July, 2006 v
Abstract This research attempted to explain the relationship among HR investment, organizational human capital and firm performance. A positive link between a firm’s HR investment and its performance through the mediating effect of overall human capital was proposed. Alternative hypotheses were presented to test the effect of different HR investment portfolios on various human capital dimensions, as well as the link between these human capital dimensions and firm performance outcomes. A review on the concept of human capital revealed several problems in the research of human capital at organizational level. Four new latent constructs (quantity of human capital, human capital-organization fit, complementarity of human capital, and specificity of human capital) were extracted from the literature to form a new paradigm in the measurement of organizational level human capital. This new paradigm represented a resource-based perspective. Data were collected from top executives of 105 companies located in the US and in Taiwan in the knowledge-intensive industry segments such as professional service, financial service, R&D, and hi-tech manufacturing, etc. Survey questionnaires were used as data collection instrument. Confirmatory factor analysis using LISREL was performed to test validity and reliability of new measurement scales. Hierarchical regression statistics were used to test the hypotheses. The results showed that HR investment had significant positive impact on firm performance and was significantly related to higher level of human capital in a firm as measured by the quantity of human capital, human capital-organization fit, complementarity of human capital, and specificity of human capital. Further, the mediating effect of firm-level human capital between HR investment and firm performance was substantiated. This study also tested a more complex model linking two HR investment portfolios to four dimensions vi
of human capital and firm performance outcomes. The findings showed that higher level of acquisition investment was linked to higher level of human capital-organization fit, complementarity of human capital, and specificity of human capital. More significantly, a higher level of development investment was linked to higher levels of all four dimensions of human capital. In addition, each individual dimension of human capital, except the quantity of it, was found to positively predict firm performance outcomes. Key words: HR investment, human resource management, human capital, firm performance, resource-based view v ii
?멋††굮??ꖻ곣ꡳꕈ룪랽냲슦왛쉉놴끑ꑈꑏ룪랽룪뭐닕슴뱨ꚸꪺꑈꑏ룪ꖻ닕슴셚껄꒧뚡ꪺ쏶ꭙꅃꖻ곣ꡳ뭻결닕슴ꪺꑈꑏ룪랽룪라덺륌닕슴ꑈꑏ룪ꖻꪺ꒤꒶껄ꩇ맯닕슴셚껄늣ꗍꖿꙖꪺ뱶암ꅃ낣ꑆ덯ꑔ귓엜볆꒧뚡쏶ꭙꪺ뻣엩껄ꩇꪺ낲뮡ꅁꖻ곣ꡳ뙩ꑀꡂ뒣ꕘꟳ닓뾰ꪺ낲뮡ꅁꕈ룕맏쓄굺꒣Ꙑꪺꑈꑏ룪랽룪닕Ꙙ녎꙰꛳뱶암꒣Ꙑ멣궱ꪺ닕슴ꑈꑏ룪ꖻ닕슴셚껄ꅃ맯ꑈꑏ룪ꖻ덯귓랧꧀ꪺꓥ쑭놴끑엣ꗜ륌ꕨꙢ닕슴뱨ꚸꪺꑈꑏ룪ꖻ곣ꡳꚳ덜Ꙩ랧꧀뿅뙱뭐ꓨꩫꑗꪺ냝썄ꅁꙝꚹꖻ곣ꡳꗑꓥ쑭�꣺ꑆꕼ귓띳ꪺ볧Ꙣ멣꧀꣓뿅뙱닕슴뱨꿅ꪺꑈꑏ룪ꖻꅁ꓀ꝏ걏ꑈꑏ룪ꖻ뙱ꅂꑈꑏ룪ꖻ빁끴ꯗꅂꑈꑏ룪ꖻ곛뮲꧊ꅂ뭐ꑈꑏ룪ꖻ녍뷨꧊ꅃꖻ곣ꡳꕈ덯귓띳ꪺ닕슴뱨ꚸꑈꑏ룪ꖻ뿅뙱볒ꚡ뙩ꛦ룪껆ꚬ뚰뭐낲덝엧쏒ꅁꕈ낪뚥ꕄ뫞뛱떪냝ꣷꝀ결룪껆ꚬ뚰ꑵ꣣ꅁ룪껆꣓ꛛ?깡과냪뭐ꕸ왗ꪾ쏑녋뚰늣띾ꪺ꒽ꕱꅁ꣒꙰녍띾?ꑵ냓?ꩁ냈ꅂ곣땯ꅂ뿄룪ꅂ낪곬빐냢ꩁ냈ꅂ뭐낪곬뭳덹띾떥ꅃ띳땯깩ꪺ멣꧀듺뙱볒ꚡꗽꕈ䱉卒䕌뙩ꛦ엧쏒꧊ꙝ꿀꓀꩒꣓듺룕걏ꝟ꣣돆ꭈꅂ껄ꯗꅁ낲뮡ꪺ엧쏒ꭨꕈ뱨꿅끪쉫뙩ꛦꅃ끪쉫엧쏒ꪺ떲ꩇ엣ꗜꑈꑏ룪랽룪뵔맪라뱶암닕슴ꑈꑏ룪ꖻꪺ뭗뽮ꅁꛓꕂꑈꑏ룪랽룪라덺륌닕슴ꑈꑏ룪ꖻꪺ꒤꒶껄ꩇ맯닕슴셚껄늣ꗍꖿ궱ꪺ뱶암ꅃꖻ곣ꡳꣃ엧쏒꣢뫘ꑈꑏ룪랽룪닕Ꙙ뭐ꕼ귓닕슴ꑈꑏ룪ꖻ멣궱꒧뚡ꅁꕈ덯ꕼ귓닕슴ꑈꑏ룪ꖻ멣궱뭐ꑔ뫘꒣Ꙑ닕슴셚껄볐꒧뚡ꪺ쏶ꭙꅃ떲ꩇ엣ꗜ닕슴Ꙣ귻ꑵ뱸뿯겡냊ꪺ룪ꕩꕈ뒣ꑈꑏ룪ꖻ빁끴ꯗꅂꑈꑏ룪ꖻ곛뮲꧊ꅂ뭐ꑈꑏ룪ꖻ녍뷨꧊ꅁꟳ엣뗛ꪺ걏닕슴Ꙣ귻ꑵ냶겡냊ꪺ룪꒣ꕩꕈꑪ둔뒣덯ꑔ귓ꑈꑏ룪ꖻ멣궱ꅁꟳ맯ꑈꑏ룪ꖻ뙱ꚳꖿ궱ꪺ뱶암ꅃꚹꕾꅁ낣ꑆꑈꑏ룪ꖻ뙱ꅁ닕슴뱨ꚸꑈꑏ룪ꖻꪺꙕ귓멣궱곒맯ꙕ뚵닕슴셚껄볐늣ꗍꖿ궱ꪺ뱶암ꅃ??쏶쇤꙲ꅇꑈꑏ룪랽룪ꅂꑈꑏ룪랽뫞뉺ꅂꑈꑏ룪ꖻꅂ닕슴셚껄ꅂ룪랽냲슦뷗 v iii
Chapter I: Introduction During the last two decades, the world has seen a dramatic shift in the business conditions in which companies compete. Many of today’s most influential firms derive their competitive advantage from sources different from the traditional conception of wealth creation a few decades ago (Pfeffer, 1998; Sveiby, 1997). Since the industrial revolution, firms have relied heavily on physical assets such as the land, natural resources, buildings, and machines, to create wealth. The emergence of knowledge-based industries and organizations (., Microsoft, .com companies) radically changed the nature in which wealth is created (Sveiby, 1997; Stewart, 2001). These firms have created value and market with few or none of the physical assets described above, but rather with their intangible assets. This change in competitive situation has resulted in a paradigm shift in business models and strategic management research in recent years. One of the most prominent strategic theories in this era is the resource-based view of the firm. The resource-based view (RBV) has shifted organizations’ strategic attention from external environment to internal resources (Priem & Butler, 2001). It has been used to extend current understanding of how resources are applied and combined, what makes competitive advantage sustainable, and the nature of rents (Collins, 2000). The characteristics of those competitive resources—as outlined in Barney (1991) as valuable, rare, difficult to imitate and non-substitutable—combined with the force of the knowledge economy to compete on intangibles, clearly established the importance of human resource management (HRM) as a source of sustainable competitive advantage for an organization (Lado & Wilson, 1994; Pfeffer, 1994; Wright & McMahan, 1992). With the support of the resource-based view, the role of HRM gained unprecedented attention and respect from the academics and 1
practitioners since the 1990s. Numerous empirical evidences also confirmed the contribution of HRM to an organization (Becker & Gerhart, 1996; Paauwe & Richardson, 1997). This new found infatuation to establish HR’s strategic position toward firm performance gave rise to a new management field aptly called “strategic human resource management” (SHRM). Rationales of the Study 1. “People are our most valuable asset.” Really? With the heightened visibility and credibility brought forth by the development of RBV and SHRM, it is only logical to expect a more powerful human resource department with better recognition, proper authority and appropriate budget; however, that does not appear to be the case in reality. With a few notable exceptions, HR departments are most often regarded as a “cost center” with expenses and overheads to be controlled, instead of being treated as a “profit center” with value propositions to be managed. Although the expression “employees are our most important assets” appears often in annual reports and press releases, field studies have shown that an “overwhelming majority of corporate executives seem to be giving lip service to the notion of people as strategic assets (Davenport, 1999, ).” Wright, Dunford & Snell (2001) observed that although HRM has established its position as strategic partner, it suffers the earliest cutbacks during tough times. As can be observed in the field, when the time gets tough, HR budget is among the first to be cut. Downsizing has always being a top-of-the-list solution to bring cost down. Some of these phenomena may have been the result of a lack of professionalism and effectiveness on the part of HR, but more often they reflect the executives’ skepticism of a proper return on HR investment (HR ROI). Without a clear HR ROI, HR managers are left with their own experience and intuitions to defend their budget. Becker, Huselid, & Ulrich (2001) suggest 2
that HR managers need to convincingly showcase HR’s impact on business bottom line to ensure HR’s strategic contribution. Pfeffer (1998) also emphasizes the need to show HR’s impact on corporate bottom line. Therefore, it is important to study HR ROI in order to establish HR’s strategic position in an organization. This study proposes measuring the relationship between HR investment and firm performance as one way to determine HR ROI. HR investment is the amount of effort placed on a meaningful set of HR practices; the returns are increased performance outcomes over the competitors. 2. A review of SHRM literature calls for more research of the black box between the HR-performance link. Examination of recent empirical studies in the strategic human resource management field has led researchers to advocate investigations of the “black box” between HR practices and firm performance (Becker & Gerhart, 1996; Wright & Gardner, 2000; Delery & Shaw, 2001). The “black box” refers to those potential intermediate links that offer causal explanation of why certain HR practices or combination of practices result in a specific level of organizational performance, or, as Wright & Gardner (2000) put it in research terms, “the intervening variables between the measure of HR practices and the measure of firm performance.” Since, researchers have conducted empirical studies to test various concepts as potential links in the HR-firm performance relationship. These include employee skills, employee behaviors, and employee motivation as shown in a review by Wright & Gardner (2000), in addition to voluntary turnover and safety as shown in the Delery & Shaw’s review in 2001. These intervening variables are, without exception, salient aspects of a firm’s human capital (HC). Thus, it is reasonable to investigate using human capital as an overarching concept of a mediating effect between HR investment and firm performance. Originally an economic terminology, the construct of human capital has evolved over time in the 3
management literature to include a broad range of human attributes that have the potential to accrue value. These human attributes, mostly possessed by an individual, include skills, knowledge, abilities (KSAs), and work motivation. Recently, researchers are beginning to look into human capital at the firm level. Some researchers use a simple aggregation of individual attributes (., Lynn, 2000; Youndt, Subramaniam, & Snell, 2004), others begin to contemplate on the complexity of human-task-organization interaction and define firm-level human capital as configurations of complementary skills and knowledge (Storey, 1995) and a firm’s ability to produce service (Pennings, Lee, & Van Witteloostuijn, 1998). So far, researchers have not agreed on an appropriate definition of human capital as it applies to organizational level research. This leads to a plethora of literature each advocating a different method for measuring the concept. This research proposes an integrative model of human capital at the firm level to help extricate the relationship between HR investment and firm performance. 3. Few attempts have been made to directly test the core concept of the resource-based theory. Researchers have increasingly relied on the resource-based view of the firm as a means of explaining why systems of HR practices lead to performance (Collins, 2000). However, Barney (2001) recognizes that most research applying the RBV has failed to test its fundamental concepts. He remarks that much of the existing research has used the RBV to “establish the context of some empirical research and are not really direct tests of the theory developed in the 1991 article (Barney, 2001, p. 46).” Wright et al. (2001) recognize that much of the existing SHRM research falls into this category. They observe that ultimately, most of the empirical studies assess only two variables: HR practices and performance. Wright et al. (2001) also point out two failures to adequately test the RBV: first, there is no attempt to empirically assess the validity of the proposition that HR practices may 4
be difficult to imitate (., are path dependent or causally ambiguous); second, although there have been attempts to demonstrate the link between HR practices and workforce characteristics, or that between the workforce characteristics and firm performance, no study has yet demonstrated “a full causal model through which HR practices are purported to impact firm performance (p. 709) .” Based on these observations, these authors suggest SHRM researchers to move beyond simply applying the “RBV logic” to HR issues and to test the “RBV’s core concepts” directly. This study attempts to address two issues described above. First, from the human capital theoretic perspective, the mediating effect of human capital is introduced to uncover one of the black boxes in the causal link between HR and performance. Second, this study proposes a direct test of the resource-based hypotheses by extracting and testing the critical characteristics of human capital in a firm that may serve as the source of sustainable competitive advantage. Research Questions As with a piece of tangible investment such as the land, a factory building or a machine, the acquisition and maintenance of human asset requires considerable monetary and non-monetary investment. Therefore, a careful assessment must be made based on returns that can be expected. The current study thus attempts to answer the following research questions: 1) Can investment in human resource accrue valuable human capital to improve company performance? 2) If so, can we predict the return on such investment like other investments made by the company? 3) How should a firm invest in HR to achieve its desired performance? 5
4) Do different types of HR investment (., buy vs. make) affect the stock of human capital in the firm? 5) Finally, how does a firm’s human capital mediate the relationship between HR investment and firm performance? These questions have enormous implications to the survival and development of HR as a profession, and yet have not been fully addressed by existing research. Research Objectives This study follows the SHRM tradition of examining a HR-performance link, and draws on the resource-based perspective as its theoretic framework to test the viability of human capital as the source of a firm’s sustainable competitive advantage. This research hopes to achieve the following objectives: 1) To prove HR’s impact on corporate bottom line by showing the relationship between HR investment and firm performance; 2) To construct a reliable and valid human capital measurement at firm level; 3) To empirically test human capital as a source of competitive advantage using the resource-based theory; and 4) To identify the relationship between different types of HR investment and their returns on human capital and firm performance. All analyses are targeted at the organizational level. The primary hypothesis is that the amount of investment on a firm’s human resource will have a positive impact on the stock of its human capital, and that a higher stock of organizational level human capital is associated with higher productivity, financial, and market performance of a firm. To further explicate the complex relationship among these three variables, two HR investment portfolios derived from the buy vs. make HR strategy dichotomy are proposed and their relationships 6
with four organizational level dimensions of human capital and three aspects of firm performance hypothesized. In Chapter II, several areas of literature that are important for understanding strategic human resource management and human capital are reviewed. The resource-based view and SHRM literature to understand how HR practices may create sustainable competitive advantage are reviewed. Further, literatures on the historical perspective, definitions, and measurements of human capital are examined to develop a new paradigm of defining and measuring organizational level human capital that takes into consideration the important interaction effects between employee attributes and organizational tasks, values, goals and demands. In Chapter III, the theoretical framework of this study is described to derive hypotheses of the proposed relationship among HR investment, firm level human capital, and firm performance. The methodology used to test these hypotheses is described in Chapter IV, which includes sections on sampling and data collection procedures, sample description, data analysis strategies, definition and measurement of variables, and scale validation procedures. Chapter V shows the analyses and results of hypothesis testing. Discussions and implications of the data analysis results are presented in Chapter VI which also includes research contributions, limitations of this study, suggestions for future research and conclusions. 7
Chapter II: Literature Review Resource-Based View of the Firm RBV is an efficiency-based explanation of performance differences and a firm-level analytical tool (Peteraf & Barney, 2003). The primary argument of RBV is that competitive advantage derives from firm-specific resources that are scarce and superior in use, relative to others (Barney, 1991; Peteraf & Barney, 2003). Although Wernerfelt’s (1984) article “A resource based view of the firm” in Strategic Management Journal signified the first coherent statement of the theory, Barney’s 1991 seminal article in Journal of Management “Firm resources and sustained competitive advantage”, in which he specified the characteristics necessary for a sustainable competitive advantage, seemed to popularize the theory within the strategy and management literatures (Wright, Dunford, & Snell, 2001). Figure 1 represents Barney’s RBV framework in this seminal article. Value Firm Resource Rareness Heterogeneity Imperfect Imitability Sustained - History Dependent Competitive Firm Resource - Causal Ambiguity Advantage Immobility - Social Complexity Substitutability Figure 1. The resource-based view of the firm Source: Barney, J., 1991, p. 112. Prior work on strategy focusing on opportunities and threats facing firms assumed that firms are homogeneous, and that any heterogeneity would quickly be dissipated. Barney 8
(1991) developed his resource-based argument using two alternative assumptions, that firms may be heterogeneous with respect to strategic resources and that those resources are not perfectly mobile. His definition of firm resources include “all assets, capabilities, organizational processes, firm attributes, information, knowledge, etc. controlled by a firm that enable a firm to conceive of and implement strategies that improve its efficiency and effectiveness.” These resources can be classified as physical capital resources, human capital resources, and organizational capital resources. Barney (1991) argues whether or not a competitive advantage is sustained depends upon the possibility of competitive duplication. He assumes that a firm has competitive advantage when it is implementing a value creating strategy that is not being implemented by any current or potential competitor. A competitive advantage is sustained if it continues to exist after efforts to duplicate it have stopped. Sustained competitive advantages may be destroyed by changes in the industry, but not by competition. He also argues that competition with homogeneous and perfectly mobile resources is not a likely scenario because under these conditions, firms cannot obtain sustained competitive advantages. Anything that one firm can do, all firms can do, including being a “first-mover”. Therefore, barriers to entry can only exist if there is heterogeneity of resources and imperfect mobility. To clearly qualify firm resources as sustained competitive advantage, Barney further explicates the properties of these resources and how these properties contribute to a firm’s competitive stand. Barney (1991) asserts that to be a potential source of sustained competitive advantage, a resource must have four attributes: it must be 1) valuable, 2) rare, 3) imperfectly imitable, and 4) not substitutable. 1) A resource must be valuable, and able to exploit opportunities or neutralize threats. When a resource is valuable, it allows a firm to conceive and implement strategies to improve efficiency and effectiveness. If a resource is not valuable, it is not a resource. 2) A resource must be rare among current and potential competitors. If everyone 9
has the resource, then no one can gain advantage from it. Valuable but common resources can help ensure survival but not competitive advantage. 3) A resource also must be imperfectly imitable. Valuable and rare resources are only sources of sustained competitive advantage if other firms cannot obtain them. A resource can be imperfectly imitable for three reasons. First, ability to obtain the resource depends on unique historical conditions. Second, the link between the resource and the sustained competitive advantage is causally ambiguous. Third, the resource is socially complex. 4) A resource must not be substitutable. In other words, there must be no strategically equivalent valuable resources that are either rare or not imitable. Strategically equivalent resources may be similar to another firm’s resource (., the top management team) or very different (., substituting a charismatic leader for a systematic company-wide planning system). This analytical framework--that organizational resources that are valuable, rare, difficult to imitate and non-substitutable (VRIS) can yield sustained competitive advantage (Barney, 1991)--has become the most cited RBV literature, despite debates and critics about the efficacy of RBV as a theory (Priem & Butler, 2001). Since Barney’s articulation of the VRIS framework, substantial effort has been devoted in the last decade to enhancing the theoretical perspectives of RBV and empirically examining it in the profit-maximizing firms (Carmeli & Tishler, 2004). The resource-based view (RBV) has become an important foundation for empirical work in strategy management and firm governance. Researchers following this approach have argued that intangible assets appear to be the key sources of strategic advantage because of their innate qualities of being more difficult to identify and imitate than tangible assets (Barney, 1991). The RBV’s shifting emphasis in the strategy literature away from external factors toward internal firm resources as sources of competitive advantage has been instrumental to the development of Strategic Human Resource Management (Wright, 1 0
Dunford, & Snell, 2001). Since Barney’s (1991) article, the RBV has become the theory most often used within SHRM, both in the development of theory and the rationale for empirical research (Wright, Dunford, & Snell, 2001). Strategic Human Resource Management Strategic human resource management is a relatively new field of research that distinguishes itself from the traditional role of human resource management as functions of personnel administration and labor relations. It also represents a new focus on HR as having value creation role, separate from prior focus on cost minimizing (Becker & Gerhart, 1996). It seeks to examine human resource management from a macro perspective, examining the effects of HR practices on firm-level performance variables instead of individual-level outcomes (McMahan, Virick, & Wright, 1999). Wright & McMahan (1992) defined SHRM as “the pattern of planned human resource deployments and activities intended to enable the firm to achieve its goals (p. 298).” Snell, Youndt, & Wright (1996) defined SHRM as “organizational systems designed to achieve sustainable competitive advantage through people (p. 62).” SHRM is strategic because SHRM researchers examine the impact of HR practices on firm performance at an organizational level of analysis, elevating HR decisions to a boardroom status. SHRM is strategic also because it focuses on the alignment of HR practices with firm strategy as a means of gaining competitive advantage (McMahan et al., 1999). SHRM has found much support in its value-creation role from the resource-based view of the firm. Becker & Gerhart (1996) believe a properly developed HR system can be an invisible asset that creates value by being embedded in the organizational system, and is difficult to imitate because of causal ambiguity and path dependency. 1 1
Best Practices vs. Best Fit Central issues of SHRM research centers around the notion of “best practices” vs. “best fit” (Paauwe & Boselie, 2003). The best practices approach refers to the search of a particular practice or a set of practices that leads to better performance regardless of a firm’s strategic and environmental contingencies (Huselid, 1995; Pfeffer, 1994). It implies an universalistic view which suggests that some disciplines are true across all types of organizations and industries. The universalistic view assumes that greater use of a particular HR practice (or set of practices) will always lead to better (or worse) performance (Delery & Doty, 1996). There has been a high level of support for these predictions. For example, Delery & Doty (1996) found three HR practices—profit sharing, results-oriented appraisals, and employment security—have generally positive effects across all firms in their study of loan officers in the banking industry. Studies by Arthur (1992, 1994), Huselid (1995), Delaney & Huselid (1996), and Ichinowski, Shaw, & Prennushi (1997) all found evidence suggesting that a “high-performance” set of HR practices do have significant impact on a variety of firm performance measures. However, there are still questions about what makes up a set of high-performance practices. Becker & Gerhart (1996) pointed out that researchers have not used a consistent set of best practices and that there is variability as to whether an individual practice should be positively or negatively related to firm performance. Becker & Gerhart (1996) suggested that the study of “best practices” should be considered at multiple levels. At high levels, “best practices” will represent the system architecture; while at lower levels, “best practices” will represent a set of specific practices which are consistent with each other and with the architecture. Two firms could have the same high-level architecture but different specific policies because of firm-specific contingencies (Becker & Gerhart, 1996). Therefore, a universal model is proposed to study HR investment practices at the highest level (., the architecture level) of a firm, 1 2
hypothesizing that a higher level of investment in HR will result in a higher level of human capital and firm performance. The fit issue leads to the study of internal consistency of practices employed in a firm, ., HR bundles, HR systems, and their alignment with firm strategy. Two fits are important for management: the external fit and the internal fit (Baird & Meshoulam, 1988). The external fit in SHRM refers to an alignment between HR practices and other aspects of the organization, especially firm strategy as a response to the external demands of the organization. External fit implies a contingency perspective and emphasizes a more complicated relationship between HR practices and performance that involves interaction of practices and external contingency variables (Delery & Doty, 1996). Examples of strategy contingencies include Miles & Snow’s (1978) strategic typology of prospector, analyzer, and defender (see Delery & Doty, 1996), and Porter’s (1985) competitive typology of cost leadership, product differentiation, and focus (see Huselid, 1995). Although the argument for external fit is compelling, the empirical work in SHRM has suffered from finding support for it. In an early review of SHRM research, Lengnick-Hall & Lengnick-Hall (1988) stated, “There is little empirical evidence to suggest that strategic HR directly influences organizational performance or competitive advantage (p. 468).” The empirical work of both Delery & Doty (1996) and Huselid (1995) found much more support for the best practices approach (high-performance HR practices that are related to performance regardless of strategy) than for the fit between HR practices and strategy. Finally, a more recent review by Wright & Snell (1998) attempting to demonstrate the benefit of fit between HR and strategy found that while some studies provided minor support for the efficacy of fit, overall the results were far from conclusive. The lack of evidence supporting external fit has led some scholars to advocate the “best practices” approach (Becker & Gerhart, 1996; Pfeffer, 1994, 1998) and to place an emphasis on internal fit. Becker & Gerhart (1996) suggested that the problems faced by a firm and the required HR alignment are more firm-specific than general 1 3
corporate strategies of cost leadership or differentiation. They proposed that the source of HR impact on performance may be “idiosyncratic contingency” where HR must be embedded and useful for solving problems specific to that firm in order to be of value. The internal alignment of practices in a complementary fashion is also known as horizontal fit at practice level. Horizontal fit is similar to Delery & Doty’s (1996) concept of configurational approach in that they are both based on patterns of practices and equifinality. Configuration implies maximally effective patterns of practices that exhibit nonlinear synergistic effects among practices (Delery & Doty, 1996). The configurational perspective leads to the study of HR practices as systems of employment relations, work practices and HR policies. This study also proposes a configurational model of HR investment, linking two distinct configurations of HR investment patterns to human capital and firm performance. These two configurations are derived from the “buy” vs. “make” HR strategies. When effectively implemented, each strategy has internal synergistic effect on maximizing a firm’s human capital. Studying HR Practices as Systems Prior empirical work in SHRM in the 80’s and early 90’s concentrated on the link between single HR practices with firm level outcomes such as turnover, productivity and corporate financial performance (See Huselid, 1995 for a more detailed review). However, Huselid (1995) pointed out that it is difficult to separate the effects of a single practice because of the high degree of multicollinearity between HR practices. Since firms normally use multiple practices simultaneously, these multiple practices may affect the same outcome variables (Huselid, 1995). Therefore, Huselid asserted that when the effects of other practices are not taken into account, the relationship between any one HR practice and an outcome may be overstated. He thus emphasized the importance of viewing HR practices from a systems 1 4
approach. Several other SHRM researchers have supported the notion of studying HR practices as systems. Wright & McMahan (1992) argued that firms must internally align HR practices with one another in a coherent system that supports each other and align these systems of practices with key organizational contingency variables. Snell et al. (1996) suggested that SHRM researchers should view HR from a systems or bundles perspective in which individual practices overlap and interact with one another. They believe that these sets of HR practices should act in concert to build skills and motivate employee behaviors, and that these skills and behaviors should align in a manner that leads to strategic advantage for the firm. Becker & Gerhart (1996) supported their notion by stating that previous emphasis on linking single HR practices with performance is inconsistent with the emphasis on internal fit in RBV and that research should focus on sets of complementary practices that form a coherent system. Along the same line, Ichniowski, Shaw, & Prennushi (1997) have empirically shown that systems of HRM practices determine productivity and quality while marginal changes in individual work practices have little effect. They concluded that “complementarity among work practices implies that the magnitude of the productivity effect of the system of HRM practices is larger than the sum of the marginal effects from adopting each practice.” (p. 309) In addition, Snell et al. (1996) suggested that there may be multiple bundles of HR practices that are concurrently operating inside organizations. There has been a great deal of empirical evidence that demonstrates the effect of various HR systems on firm performance (Arthur, 1992, 1994; Becker & Gerhart, 1996; Huselid, 1995; Ichniowski et al., 1997; MacDuffie, 1995; Youndt et al., 1996). For instance, Arthur (1994) found in 30 steel minimills that those with “commitment” human resource systems, emphasizing the development of employee commitment, had lower turnover and scrap rates and higher productivity than firms with “control” systems, emphasizing efficiency and the reduction of labor costs. Commitment maximizing system includes the practice of 1 5
broadly defined jobs, high employee influence and communication, high skill requirements, training, and higher wages. MacDuffie (1995) found that “bundles” of internally consistent HRM practices were associated with higher productivity and quality in 62 automotive assembly plants. Huselid (1995) studied 13 HR practices in a “High Performance Work System” (HPWS) in a large scale cross-sectional study and found HPWS to be significantly related to turnover, productivity and financial performance. Delery & Doty (1996), in a study of loan officers in the banking industry, tested the configurational perspective using “market-type” and “internal employment” systems and found a positive relationship between the market-type configuration and performance. The market-type configuration includes a higher level degree of use in results-oriented appraisals, profit sharing, and participation. Ichniowski, Shaw, & Prennushi (1997), using data from 36 steel plants, found the impact of a set of “innovative” work practices to have a positive and significant effect on organizational productivity. These innovative work practices included incentive pay, teams, flexible job assignments, employment security, and training. So far, there is no agreement as to which individual practices should be included in a system, and researchers have used many different practices to represent the same type of HR system of practices (Becker & Gerhart, 1996). This may be the result of both a lack of solid theoretical framework to guide the selection of practices in a system and the fact that many HR practices may result in the same outcomes (Collins, 2000). Another problem with the study of HR practices as systems is the issue of whether the set of practices in a system contain a multiplicative effect or an additive effect. Delaney & Huselid (1996) tested the complementarities among HR practices by analyzing the interaction effects of all possible combinations of HR practices on perceived organizational performance. They failed to find any positive effects that were derived from a specific combination of practices. Alternatively, Guest (1997) suggested that the pattern of combinations among HR practices is additive rather than multiplicative, and that the sum of each HR practice is greater than its parts. This 1 6
notion is supported by many researchers who studied HR as a system using additive measures (., MacDuffie, 1995; Delery & Doty, 1996). Following Guest (1997), the two HR investment configurations in this study are developed using additive measures. Limitations of RBV-based SHRM Empirical Research Although multiple theories have been put forth to explain how HR practices lead to higher firm performance by influencing firm resources or employee behaviors (., behavioral theory, human capital theory, etc. See McMahan et al., 1999), researchers have increasingly relied on the resource-based view of the firm as a means of explaining why systems of HR practices lead to performance (Collins, 2000). However, applying resource-based view in SHRM empirical research has its limitations. In response to critics that the RBV does not generate testable hypotheses, Barney (2001) recognizes that most research applying the RBV has failed to test its fundamental concepts. He remarks that much of the existing research has used the RBV to “establish the context of some empirical research and are not really direct tests of the theory developed in the 1991 article (Barney, 2001, p. 46) .” Wright et al. (2001) recognize that much of the existing SHRM research falls into this category. Although the underlying logic in most SHRM empirical studies—that HR activities lead to a competent, motivated workforce forming a source of competitive advantage which generates higher performance—is compelling, Wright et al. (2001) note that ultimately, most of the empirical studies assess only two variables: HR practices and performance. Wright et al. (2001) also point out two failures to adequately test the RBV: first, there is no attempt to empirically assess the validity of the proposition that HR practices are difficult to imitate (., are path dependent or causally ambiguous); second, although there have been attempts to demonstrate the link between HR practices and workforce 1 7
characteristics, or that between the workforce characteristics and firm performance, no study has yet demonstrated “a full causal model through which HR practices are purported to impact firm performance (p. 709) .” Based on these observations, these authors suggest SHRM researchers to move beyond simply applying the “RBV logic” to HR issues and to test the “RBV’s core concepts” directly. This calls for a more complex view of the relationship between HR and performance within the empirical literature, for example, recognizing the black boxes in the causal link, specifying the role HR practices play in developing competitive advantages of firms, etc. This study attempts to address two issues described above. First, from the human capital theoretic perspective, the mediating effect of human capital is introduced to uncover one of the black boxes in the causal link between HR and performance. Second, this study proposes a direct test of the resource-based hypotheses by extracting and testing the critical characteristics of human capital in a firm that may serve as the source of sustainable competitive advantage. Human capital is reviewed in the next section as a black box between HR and firm performance. Human Capital as a Black Box between HRM and Firm Performance Human capital is a very loaded term. A half-century history of rich research application and practitioner attention has added layers upon layers of meanings to the concept. To discern the laden meanings of human capital requires an understanding of its historical perspective, its current development in management study, and a clarity on the level of analysis. This section is thus organized to describe 1) the history and application of the HC concept; 2) the definition of HC; 3) the measurement of HC; and 4) measuring HC for organizational level analyses. 1 8
Development and Application of the Human Capital Concept Mincer (1993), who have studied human capital for nearly a half century, quoted Irving Fisher’s definition of capital as any asset that gives rise to an income stream and stated that accumulated human work capacity qualifies as a capital asset. The concept of human capital has a long history. Mincer (1993) believes the development of human capital theory appeared as a response to empirical findings observed by growth theorists in the 1950s. Technical change, quality of capital and quality of labor (., human capital) were implicated in the large growth residual when the observed growth of conventionally measured inputs of labor and capital fell behind the growth of output in the US and in other countries over long periods of time (Mincer, 1993). The term human capital was first introduced by Nobel Prize-winning economist Theodore W. Schultz in a 1961 American Economic Review article: “Investment in Human Capital.” Schultz popularized the concept of human capital by emphasizing the importance of investments in human productivity in economic growth. These investments include education, job training, health, and migration in human capital. At the same time, Mincer started a series of studies exploring the use of amounts of training and experience on the job to explain differences in earnings, and the costs and a rate of return on on-the-job training (1993). The analytical power of human capital became explicit in Gary Becker’s classic text, Human Capital (1964, first presented at a National Bureau of Economic Research conference in 1961), where a comprehensive human capital investment theory can be found. His model used job training as a framework in which to calculate the investment in human capital of both individual and firms and distinguished between general and specific human capital. He hypothesized that employees with specific human capital will have a lower turnover rate (Becker, 1962). Becker’s work represents the mainstream of human capital theory particularly with regard to specific human capital (Lacey, 1982). 1 9
The theory of human capital was traditionally used by economists to analyze the effects of human capital at the macroeconomic level. The focus of the investigation was the relationship between human capital and economic outcomes at the social or national level such as economic activities, societal well-being, and growth of per-capita income (Nübler, 1997). The main subject of the initial work in human capital focused on the explication of costs and returns to education and other investments in human capital at a private or social level. Subsequently, human capital theory was widely applied to the study of the structure of personal earnings, by schooling, labor market experience, occupation, sex, region, and so forth (Mincer, 1979). The concept of human capital became important in the business-level analysis when companies, driven by the forces of the economy, ran out of ammunitions on the more conventional competitive advantage and began to compete on their intangibles (Barney & Wright, 1998; Becker & Huselid, 1998). The study of the knowledge-based economy and intellectual capital in the 1980s and 1990s also brings human capital to the fore front of management research. Human capital, regarded as either a critical driving force, a foundation, a pillar, or a repository of a firm’s intellectual capital (Edvinsson & Malone, 1997; Stewart, 1997; Sveiby, 1997), has become one of the most researched subject in the knowledge management literature. Definition of Human Capital Human capital is an intangible resource and a somewhat elusive concept, leading researchers to view and define it in a number of different ways. See Table 1 for a listing of human capital definitions by various authors. As the first author who coined the term “human capital’, Schultz (1961) used it to refer to the health, skills and knowledge of an individual which have economic value. Becker (1964) further defined human capital as the knowledge, skills, and abilities of employees in a firm’s workforce. Most researchers followed or 2 0
extended from this definition to include constructs that believed to have contributed to personal productive capacity, such as education, training, and experience (., Bontis & Fitz-enz, 2002; Crawford, 1991; Lepak & Snell, 1999; Rosen, 1987; Youndt, Subramaniam, & Snell, 2004). Some authors, influenced by the field of I/O psychology, added attitudinal dimensions such as work motivation and commitment to the definition of HC (., Edvinsson & Malone, 1997; Nordhaug, 1993; Sveiby, 1997). Nordhaug’s (1993) definition of human capital distinguishes between an individual’s potential ability to do a job and his/her willingness to perform. Health and competences constitute the individual employee’s basic capacity to perform tasks; while work motivation and commitment influence the actual performance. Nordhaug believes these four elements of human capital define the individual employee’s capability in work. Sveiby (1997) believes that value judgments are a component of competence, although he did not place value and attitude of employee as an indicator in the measurement of competence, but rather under the measurement of internal structure efficiency. He believes the kind of attitude that deserves tracking is often referred to as corporate culture or esprit de corps, which contributes consciously or unconsciously to employees’ behavior in enhancing or breaking the company’s image in front of its customers. In conclusion, human capital of an individual employee can be defined as his/her productive capacity at work which may be determined by his/her skills, knowledge, abilities, attitudes, and health. 2 1
Table 1. Definitions of human capital Reference Definition Level of Analysis* Schultz, 1961 health, skills and knowledge, which have economic value Ind. Becker, 1964 the knowledge, skills, and abilities of employees in a firm’s workforce Ind. Flamholtz, 1972 productivity, transferability and promotability Ind. Rosen, 1987, p. 681 "... the productive capacities of human beings as income producing Ind. agents in the economy” Crawford, 1991, p. 10 the knowledge, education, training, skills, and expertise of a firm’s Ind. workers Hudson, 1993 your genetic inheritance; your education; your experience; and your Ind. attitudes about life and business Becker, 1993 the knowledge, skills, abilities, personality, appearance, reputation, and Ind. credentials Nordhaug, 1993 health, competences (KSAs that are relevant for work), work Ind. motivation and commitment Storey, 1995, p. 4 intangible assets such as unique configurations of complementary skills, Org. and tacit knowledge, painstakingly accumulated, of customer wants and internal processes Brooking, 1996 skills and expertise, problem-solving abilities, leadership styles and Ind. abilities and everything that is embodied in the employees Saint-Onge, 1996 is embodied in the aptitude of individuals in the workplace to determine Ind. the best options or solutions for clients Edvinsson & Malone, the ability, knowledge, skills, innovation and experiences of employees Ind./Org. 1997 and managers, also includes the firm’s values, culture, and philosophy Stewart, 1997, 2001 talent (the skills and knowledge of employees), individual ability to Ind. solve customer problems, a source of innovation and renewal Sveiby, 1997 employee competence which involves the capacity to act in a wide Ind. variety of situations to create both tangible and intangible assets Pennings, Lee, & Van a firm’s ability to produce a high-quality service, the knowledge and Org. Witteloostuijn, 1998 skills of its professionals that can be used to produce professional services Roos, Roos, Edvinsson, competence, attitude and intellectual agility Ind. & Dragonetti, 1998 Davenport, 1999 ability, behavior, effort and time Ind. Lepak & Snell, 1999 the skill, ability of employees in the organization which make potential Ind. contribution to the organization’s competitive advantage and core competency Lynn, 2000 all skills and capabilities of the people working in an organization, an Org. inventory of individual’s skills and knowledge within an organization Sullivan, 2000 capabilities of employees to solve customer problems, collective Org. experience, skills, and general know-how of all of the firm’s people Bontis & Fitz-enz, the sheer intelligence of the organizational member, the knowledge, Ind. 2002 talent and experience of employees Piazza-Georgi, 2002 the stock of personal skills that economic agents have at their disposal Org. Kocakulah & Harris, the people assets based on the unique capabilities and expertise of each Org. 2002 individual and collectively to the organization Chen & Lin, 2003 investments made by company in talents and technologies that benefit Org. competitive advantage, are valuable and unique, and should be kept out of reach of other companies Youndt, Subramaniam, the knowledge, skills, and abilities (KSAs) employees possess that Ind. & Snell, 2004, p. 345 bring economic value to firms * Ind.: individual level. Org.: organizational level. Source: compiled for this study 2 2
When the term human capital was first created, it was used largely as an individual level construct in the analysis of the effect of investment and return on people (Schultz, 1961). Thus, definitions of HC include mostly constructs that are espoused from the perspective of an individual level analysis such as knowledge, skills, abilities, experience, and motivation (see Table 2 for a classification of constructs used to define HC). Although HC was later adopted by the knowledge management theorists as an organizational level concept to represent the overall strength of a firm’s workforce to produce intellectual capital (Davenport, 1999; Edvinsson & Malone, 1997; Stewart, 1997, 2001), the individual perspective still pervades the definition and measurement of this concept. Aggregation of employee KSAs is the most commonly seen definition of HC when referring to a firm’s human capital (See Bontis & Fitz-enz, 2002; Brooking, 1996; Crawford, 1991; Lepak & Snell, 1999; Lynn, 2000; Saint-Onge, 1996; Sullivan, 2000). Aggregation implies the concept of quantity that can be expressed as “amount” or “level”. Using aggregation of individual human capital attributes such as employee KSAs as the definition of firm level human capital inevitably precipitates the use of quantity measures such as the number of people in possession of certain KSAs or the average level of KSAs employees possess. A very limited number of writers have extended the definition of firm level HC beyond a simple aggregation of employee KSAs. They proposed additional criteria in the definition of firm level human capital such as customer demand, collectivity, uniqueness, and value to the organization. Storey (1995), from a resource-based view, described human capital as intangible assets resulted from “unique configurations of complementary skills, and tacit knowledge, painstakingly accumulated, of customer wants and internal processes (p. 4).” Edvinsson & Malone (1997) included a firm’s values, culture, and philosophy to the definition of HC, in addition to the ability, knowledge, skills, innovation and experiences of employees and managers. Sullivan (2000) defined human capital as capabilities of employees to solve 2 3
customer problems, and the collective experience, skills, and general know-how of all of the firm’s people. Kocakulah & Harris (2002) posited that human capital is the people assets based on the unique capabilities and expertise of each individual and collectively to the organization. Chen & Lin (2003) defined human capital as the investments made by company in talents and technologies that benefit competitive advantage, are valuable and unique, and should be kept out of reach of other companies. Pennings, Lee, & Van Witteloostuijn (1998) focused on a firm’s ability to produce a high-quality service and defined HC as the knowledge and skills of its professionals that can be used to produce professional services. This later group of writers carries a more selective view of human capital in a firm. They do not believe that every individual KSA or attribute presents value to the organization, even if an attribute has proven to be valuable to the person carries it. Only those attributes that can be combined to satisfy customer demands and to produce economic value to the firm qualify as the firm’s human capital. As firms handle very different customer demands or use different tactics to handle these demands, the composition of organizational level human capital also differ greatly across firms. The composition differs not only in the inventory of individual human capital in a firm, but also in the way these individual human capital are combined to form collective strength that matches the design of the organization in producing products and services to satisfy customer needs. Synthesizing these definitions in the literature, a more comprehensive definition of the organizational level human capital is proposed as “the level and the unique configurations of complementary employee attributes that fit the organizational value and goals in fulfilling customer demands in order to bring economic value to the firm.” This definition conveys not only the quantity, but also the HC-organization fit, the complementarity and the uniqueness of a firm’s human capital. 2 4
Table 2. Classification of human capital constructs Level of Construct Reference Analysis* Ind. Competencies/capabilities/abilities: Becker, 1964; Brooking, 1996; Davenport, 1999; Edvinsson & Malone, 1997; Lado & Wilson, 1994; Lepak & Snell, 1999; Lynn, 2000; Nordhaug, 1993; Roos et al., 1998; Saint-Onge, 1996; Schultz, 1961; Stewart, 1997; Sveiby, 1997; Youndt, Subramaniam & Snell, 2004 Ind. Skills Becker, 1964; Brooking, 1996; Crawford, 1991; Edvinsson & Malone, 1997; Lado & Wilson, 1994; Lepak & Snell, 1999; Lynn, 2000; Nordhaug, 1993; Pennings, Lee, & Van Witteloostuijn, 1998; Schultz, 1961; Stewart, 1997, 2001; Youndt, Subramaniam & Snell, 2004 Ind. Knowledge Becker, 1964; Crawford, 1991; Bontis & Fitz-enz, 2002; Edvinsson & Malone, 1997; Lado & Wilson, 1994; Nordhaug, 1993; Pennings, Lee, & Van Witteloostuijn, 1998; Schultz, 1961; Stewart, 1997, 2001; Youndt, Subramaniam & Snell, 2004 Ind. Experience Bontis & Fitz-enz, 2002; Edvinsson & Malone, 1997; Hudson, 1993 Ind. Attitude/value/belief/mindset Hudson, 1993; Nordhaug, 1993; Roos et al., 1998 Ind. Know-how/expertise Brooking, 1996; Crawford, 1991 Ind. Health Nübler, 1997; Schultz, 1961 Ind. Innovation Edvinsson & Malone, 1997 Ind. Productivity Flamholtz, 1972; Rosen, 1987 Ind. Education Crawford, 1991; Hudson, 1993 Ind. Training Crawford, 1991 Ind. Intelligence Bontis & Fitz-enz, 2002 Ind. Talent Bontis & Fitz-enz, 2002 Ind. Behavior Davenport, 1999 Ind. Effort Davenport, 1999 Ind. Time Davenport, 1999 Ind. Intellectual agility Roos et al., 1998 Ind. Genetic inheritance/attributes Hudson, 1993 Ind. Transferability Flamholtz, 1972 Ind. Promotability Flamholtz, 1972 Ind. Commitment Nordhaug, 1993 Ind. Work motivation Nordhaug, 1993 Org. Collective capabilities and expertise Kocakulah & Harris, 2002 Pennings, Lee, & Van Witteloostuijn, 1998; Sullivan, 2000 Org. Stock of personal skills Piazza-Georgi, 2002 Org. Unique configurations of Storey, 1995 complementary skills Org. Tacit knowledge of customer wants Storey, 1995 and internal processes Org. Valuable and unique talents Chen & Lin, 2003 Org. Organizational value Edvinsson & Malone, 1997 Org. Organizational culture Edvinsson & Malone, 1997 Org. Organizational philosophy Edvinsson & Malone, 1997 * Ind.: individual level. Org.: organizational level. Source: compiled for this study 2 5
Measurement of Human Capital The history of HC measurement at an organizational level began as research in human resource costing and accounting brought forward some emphasis on personnel accounting calculations for managerial decision making (Sveiby, 1997). A wide-spread frenzy which started in the 80s and blossomed well into the 90s to study and manage intellectual capital propelled a massive collection of effort in measuring human capital at firm level (., Edvinsson & Malone, 1997; Stewart, 1997, 2001; Sveiby, 1997). Researchers have used an array of different measures of human capital in various combinations in their empirical studies designating human capital as one of the variables. For example, in Becker’s (1962) early work on a theoretical model of the investment in human capital and its marginal productivity, human capital was categorized as schooling and on-the-job training. In a 2001 study conducted by Hitt, Bierman, Shimizu, & Kochhar examining the direct and moderating effects of human capital on professional service firm performance, they used the following measures as proxies to two important dimensions of human capital: 1) quality of the law school attended by partners which served as a proxy for “articulable knowledge and prestige” and 2) total experience as partners in the focal firm which served as a proxy for “firm-specific tacit knowledge”. Pennings, Lee, & Van Witteloostuijn (1998) used proportion of partners and associates possessing master’s or higher degrees and average years of industry experience as proxies of “partners’ & associates’ industry-specific HC”, and natural log of their tenures at the firm as a proxy of “partners’ & associates’ firm-specific HC” when studying the effect of human capital on firm dissolution. Walker (2001), investigating the link between the value of human capital and firm performance, calculated the value of HC by total compensation of a firm divided by its number of employees. Carmeli & Tishler (2004) and Youndt, Subramaniam, & Snell (2004) both studied 2 6
the effect of human capital on organizational performance using perceptual questionnaire items to identify the level of skills, knowledge and competence of the workforce. The above review on empirical research using human capital as the independent variable revealed some common threads. The first revelation is the widespread use of proxies in measuring human capital, particularly the popularity of using education and tenure as substitutes for actual level of employee KSAs. Secondly, although the analysis of the above studies was set at the organizational level, researchers tended to use measures at the individual employee level such as education and tenure. Aggregation through adding or averaging those individual employee attributes was then used for organizational level analysis. Lastly, the measure of human capital in an organization was overly simplified so that only one or two dimensions of human capital (., skills and knowledge) were represented in those studies. These common threads in existing empirical research greatly relegated the viability and usefulness of human capital research at macro organizational level. In search for a more extensive coverage of organizational level human capital measurement, I then turn to the large volume of books discussing the management of intellectual capital. The work of Edvinsson & Malone (1997), Sveiby (1997) and Stewart (2001) are highlighted here. (See Table 3.) Edvinsson & Malone in their 1997 book on intellectual capital presented a universal measurement and reporting system on human capital. Their effort resulted in 23 measures including leadership, motivation, empowerment, number of employees (full-time permanent vs. full-time temporary vs. part-time/contractors), employee turnover, average employee tenure, number of managers/women managers, average age of employees, time in training, IT-literacy of staff, and education background of managers. Sveiby (1997), a pioneer in the measurement of intellectual capital, devised a framework for measuring intangible assets which he calls the Intangible Assets Monitor (IAM). The IAM is designed to monitor three intangible assets in an organization—the competencies 2 7
of its people, its internal structure, and its external structure. Under each intangible asset Sveiby includes three groups of indicators to measure the growth and renewal, the efficiency, and the stability of that asset. Sveiby believes “employee competence is not only one of the three intangible assets but also a source of the internal and external structures.” Reflecting his strong belief in human capital, all of his competence indicators and part of his internal structure indicators involve the measurement of personnel attributes. These are number of years in the profession, level of education, training and education costs, grading, professional staff turnover, competence-enhancing customers, proportion of professionals in the company, the leverage effect, value added per professional, average age, seniority, relative pay position, professional turnover rate, proportion of support staff, sales per support person, values and attitude measurements, support staff turnover, the rookie ratio. Stewart (1997) focusing on the measurement of the efficiency of human capital used more outcome-based measures in relation to the aspect of innovation. His measures of firm level HC include: average years of service, average education level, % with advanced degrees, hiring cost, IT literacy, hours of training/employee, employee satisfaction, success of employee-suggestion programs, employee turnover (separations), value-added/employee, new colleague-to-colleague relationships spawned, and various measures of innovation such as sales from new products, gross margin from new products, R&D intensity, R&D efficiency (profits from new products/R&D spending). The above review of the popular literature on the measurement of firm level human capital reveals additional problems. The use of proxies to replace measures of the actual level of human capital constructs still persists. For example, to measure employee KSAs, Edvinsson & Malone used indicators such as average years of service with company, average age, percentage of company managers with advanced degrees; Sveiby used indicators such as number of years in the profession, level of education, training and education costs, grading, age, seniority; and Stewart used indicators such as average years of service, average education 2 8
level, % with advanced degrees, hours of training/employee, etc. Aggregation from individual level to firm level human capital is still a common method, although some organizational level indicators such as leadership, motivation, empowerment, turnover, innovation, and employee relationship have started to emerge. The problem of over-simplification is reversed in the popular literature. The three highlighted human capital measurement systems are very extensive in terms of number of measures. Although the authors claimed that these measures were intended to capture the multiple facets of a firm’s human capital, they actually went beyond their definition of human capital and began to track other aspects of the organization. As categorized in Table 3, indicators such as empowerment, amount of training, having competence-enhancing customers, hiring cost, success of employee-suggestion programs, and employee relationship do not measure human capital, but rather the antecedent of human capital. On the contrary, employee turnover, employee value added, sales, and innovation are measures of the output or the outcome of human capital. Some of these are commonly seen organizational effectiveness indexes; some qualify more as HR efficiency indexes. This leads to great confusion as to what human capital is in an organization. The most serious problem in these three popular HC measurement systems is that they are fixated on a narrow definition of HC as the collection of KSAs and attitude of employees in a firm. Although their measures appear to be extensive, they fail to include important attributes such as health of employees and the complexity of matching unique configurations of complementary employee attributes to organizational needs. Health of employees, both physical and psychological, is an important factor that affects employee motivation and productivity on the job. The complexity of fit, complementarity and specificity of a firm’s collective human capital has serious implications on the firm’s capability to satisfy customer demands. A firm-level human capital measurement can not be considered complete without these important dimensions. 2 9
Table 3. Popular human capital measurements Author Indicator Nature of Indicator Type of Indicator Edvinsson & Malone, 1. Leadership index HC HC stock index 1997—Human focus 2. Motivation index HC HC stock index 3. Empowerment index Antecedent of HC Organizational effectiveness index 4. Number of employees HC HC stock index 5. Employee turnover Outcome of HC HR efficiency index 6. Average years of service Proxy for employee KSAs HC stock index with company 7. Number of managers HC HC stock index 8. Number of women HC HC stock index managers 9. Average age of employees Proxy for employee KSAs HC stock index 10. Time in training (days/year) Antecedent of HC HR efficiency index 11. IT-literacy of staff HC HC stock index 12. Number of HC HC stock index full-time/permanent employees 13. Average age of Proxy for employee KSAs HC stock index full-time/permanent employees 14. Average years with Proxy for employee KSAs HC stock index company of full-time permanent employees 15. Annual turnover of full-time Outcome of HC HR efficiency index permanent employees 16. Per capita annual cost of Antecedent of HC HR efficiency index training, communication, and support programs for full-time/permanent employees 17. Full-time/permanent Antecedent of HC HR efficiency index employees who spend less than 50 % of work hours at a corporate facility 18. Number of full-time HC HC stock index temporary employees 19. Average years with Proxy for employee KSAs HC stock index company of full-time temporary employees 20. Per capita annual cost of Antecedent of HC HR efficiency index training and support programs for full-time temporary employees 21. Number of part-time HC HC stock index employees/non-full-time contractors 22. Average duration of HC HC stock index contract 23. Percentage of company Proxy for employee KSAs HC stock index managers with advanced degrees 3 0
Table 3 continued. Author Indicator Nature of Indicator Type of Indicator Sveiby, 1997— Indicators of Growth/Renewal Measuring competence 1. Number of years in the Proxy for employee KSAs HC stock index profession 2. Level of education Proxy for employee KSAs HC stock index 3. Training and education costs Proxy for employee KSAs HC stock index 4. Grading Proxy for employee KSAs HC stock index 5. Turnover Outcome of HC HR efficiency index 6. Competence-enhancing Antecedent of HC Organizational customers effectiveness index Indicators of Efficiency 1. Proportion of professionals in Quantity of HC HC stock index the company 2. The leverage effect Antecedent of HC Organizational effectiveness index 3. Value added per professional Outcome of HC Organizational effectiveness index Indicators of Stability 1. Average age Proxy for employee KSAs HC stock index 2. Seniority Proxy for employee KSAs HC stock index 3. Relative pay position Proxy for employee KSAs HC stock index 4. Professional turnover rate Outcome of HC HR efficiency index Sveiby, 1997— Indicators of Efficiency Excerpt from measuring 1. Proportion of support staff Quantity of HC HC stock index internal structure 2. Sales per support person Outcome of HC Organizational effectiveness index 3. Values and attitude Attitudinal measure of HC HC stock index measurements Indicators of Stability 1. Support staff turnover Outcome of HC HR efficiency index 2. The rookie ratio Antecedent of HC Organizational effectiveness index Stewart, 2001— 1. Avg. years of service Proxy for employee KSAs HC stock index Measuring efficiency of 2. Avg. education level Proxy for employee KSAs HC stock index human capital 3. % with advanced degrees Proxy for employee KSAs HC stock index 4. Hiring cost Antecedent of HC HR efficiency index 5. IT literacy HC HC stock index 6. Hours of training/employee Proxy for employee KSAs HC stock index 7. Employee satisfaction Proxy for employee attitude HC stock index 8. Success of Antecedent of HC HR efficiency index employee-suggestion programs 9. Employee turnover Outcome of HC HR efficiency index (separations) 10. Value-added/employee Outcome of HC Organizational effectiveness index 11. Various measures of Outcome of HC Organizational innovation: sales from new effectiveness index product, R&D intensity or efficiency 12. New colleague-to-colleague Antecedent of HC Organizational relationships spawned effectiveness index Source: compiled for this study. 3 1
Measuring Human Capital for Organizational Level Analyses There is a large discrepancy between the definition of HC and the measurement of HC. For instance, skills and knowledge (., Becker, 1964; Bontis & Fitz-enz, 2002; Edvinsson & Malone, 1997; Youndt, Subramaniam, & Snell, 2004) are the most common definition of human capital, but are often measured using imprecise proxies such as years of education, hours of training, and tenure. Attitude (Hudson, 1993; Nordhaug, 1993; Roos et al., 1998) and health (Nübler, 1997; Schultz, 1961) are also referred to in many writings as dimensions of individual human capital, but are rarely accounted for in the measurement of either individual or organizational level human capital. Moreover, the transition of measuring human capital from an individual to an organizational perspective fails to acknowledge important attributes of fit, uniqueness and complementarity in a firm’s collective human capital. The discrepancy is likely to originate from a lack of good theoretical development in the academic world on the measurement of organizational level HC, compounded with practitioners’ eagerness to manage HC in an organization (Sveiby, 1997). This discrepancy can also be observed in academic research. With the increasing importance of intangible assets in the knowledge economy, intellectual capital and its related terms such as human capital gained unprecedented popularity in the management field. Many researchers use the term human capital loosely to mean very different constructs in their research. Some use it to alternate with the more traditional-toned term “human resource”. This phenomenon adds to the confusion of the terminology. Measurement of organizational human capital thus far is still evolving. Although there were attempts to put the various measures into perspective through some sorts of classifying scheme, there is still no agreement on how best to organize these measures. This reflects a need to better understand what consists of human capital in a firm. There are a number of issues with the research on human capital that must be addressed in order to understand how HC can 3 2
affect firm performance. First, it is important to identify the appropriate level of analysis at which this construct should be studied to determine its relationship to firm performance. Second, it is empirical to specify the nature of the measures (., antecedent, stock, or outcome of HC) so it is clear what is being measured. Third, it is critical to identify the most salient components of human capital (., sub-constructs of HC) that are meaningful when measuring HC at firm level. Level of Analysis Specifying the level of analysis is important in management research. The confusion about the appropriate level of analysis may affect the size and robustness of the relationship between independent and outcome variables of interest (Delery, 1998). The majority of current existing measure (or measures) of human capital represents an aggregated static view of the subject, ., a stock or an inventory of individual human capital at the time (., Lynn, 2000). Taking a stock or inventory means measuring attributes at the individual level then simply adding the numbers together. For example, a common measure of HC in a firm is the headcount or percentage of employees at a certain skill level, using the amount of training or experience as a proxy for skill level (See Hitt et al., 2001; Pennings et al., 1998). Aggregation from individual HC has two implications to the level of analysis in linking HC to firm performance. First, it represents a quantitative view of human capital irregardless of any contingency factors and assumes that individuals with the same personal credentials possess similar productive capacity and thus contribute equally to firm performance in each of their respective firms. The quantitative view is problematic in that it disrespects the context or the external environment of a job placement. For example, is company A better than company B because company A possesses more managers who have longer industry experience or are better educated? If the human capital possessed by company A’s managers do not satisfy company A’s strategic demand, the high level of individual 3 3
human capital will simply be a waste. Second, when individual HC is aggregated to group or organization level, the interaction effect between individual HC and task assignment, and the interaction effect between individual HC and group characteristics are not counted. Colbert (2004) presented a complex, living-systems extension of the resource-based view in strategic human resource management, which emphasizes the interaction effects and system-level characteristics of possible sources of competitive advantage. He defined system-level resources as “those organizational capabilities that exist only in relationships—in the interactions between things (Colbert, 2004, p. 348).” Therefore, a simple aggregation of individual level HC is not a complete measure to study the effect of human capital as a source of competitive advantage. Organizational level HC implies other sub-constructs to account for these interaction effects and thus requires a different paradigm in measurement. Nature of HC measures Current existing measurement of HC mix-match measures in different natures. As shown in Table 3, there are at least three schools of measures used by previous researchers and practitioners: antecedent of HC, HC stock, and outcome of HC. Those measures that are marked as antecedent of HC (., hiring cost, success of employee suggestion programs) or outcome of HC (., turnover, employee value-added) are really HR efficiency or organization effectiveness measures. They may give management an indication of how well the company is doing in the effort to increase human capital, but they do not accurately reflect the level (both in quantity and quality) of human capital currently possessed by the company. When researchers use predictors of HC to measure the stock of human capital as an outcome variable, the positive causal effect of antecedent variables such as HR investment may be biased because of confounding problems. The common practice of using proxies in HC measurement presents other problems. 3 4
For example, Chan (2000) questioned the use of job tenure as a proxy for employee experience, because empirical evidence has indicated that different individuals with the same job tenure can differ greatly in the number and type of tasks they perform. Delery & Shaw (2001) questioned using years of education as a proxy for employee KSAs. Their suspicion is quite obvious because it is a well-known fact that people with the same amount of education or even the same quality of education differ greatly in their capabilities and achievement. This problem is aggravated when imprecise proxies are aggregated to measure the stock of human capital in an organization. Measuring the whole of work force characteristics is a very difficult task (Delery & Shaw, 2001). A lack of theoretical development may explain some of the limitations of current measurement criticized by researchers. These critics brought up the issue of measuring what’s really required by the job vs. what’s the actual level in the work force (Delery & Shaw, 2001), which is the concept of fit vs. quantity. Sub-constructs of Organizational Level Human Capital Four sub-constructs (or dimensions) emerged from the above review of the literature on organizational level human capital. These are 1) quantity of human capital, 2) HC-organization fit, 3) complementarity among human capital, and 4) specificity of human capital. These four dimensions mirror roughly to the four criteria of a competitive resource specified by Barney (1991) in the resource-based view: scarcity, value, inimitability, and non-substitutability. Colbert (2004) reviewed conceptual and empirical work on RBV and ascertained the importance of “the relationships between and among resources that display ‘cogency,’ ‘complementarity,’ or ‘cospecialization’ or that generate rents at the system level or organization level (p. 348).” He believes such resources is key to sustaining economic rent since they are “system specific” (or firm specific). His discussion confirms the validity of the three new sub-constructs (., fit, complementarity, and specificity) of organizational level 3 5
human capital in my review, although he seems to assume resource specificity as a function of fit and complementarity whereas I propose specificity as a separate sub-construct. My definition of organizational level human capital as “the level and the unique configurations of complementary employee attributes that fit the organizational value and goals in fulfilling customer demands in order to bring economic value to the firm” also implies these four separate sub-constructs. Table 4 compares organizational level with individual level constructs of human capital proposed in this study. Table 4. Human capital constructs: organizational level vs. individual level Organizational level construct Individual level construct Quantity of HC Knowledge HC-organization Fit Skills Complementarity of HC Abilities Specificity of HC Attitudes Health Quantity of HC. The quantity issue is fundamental within the firm from a manpower standpoint. Firms need to employ sufficient amount of manpower with appropriate level of skills and capabilities in order to serve their customers in a timely fashion. (Michie & West, 2005) Previous literature on the measurement of firm level human capital reflects mostly this perspective, as evidenced by items asking for a headcount of employees and aggregations of employee skills and capabilities (see, for example, Edvinsson & Malone, 1997). However, according to the resource-based hypothesis, the type of human capital should be rare in the market for human capital to become sustainable competitive advantage for a firm (Barney, 1991). Levels of employee attitudes and health also have direct impact on workforce productivity. Work motivation and commitment are often cited as having great influence on employees’ willingness to perform and thus have significant consequences on their actual performance (Edvinsson & Malone, 1997; Nordhaug, 1993; Pfeffer, 1994, 1998). 3 6
Good psychological and physical health constitutes employees’ basic capacity to perform tasks (Michie & West, 2005; Nordhaug, 1993). Although physical health is rarely attended to in the personnel management literature (with the exception of studies on job safety), psychological health of employees is an important topic in I/O psychology and personnel management as evidenced by the abundant literature on job stress, psychological strains, and burnout. HC-Organization Fit. The value of a firm’s human capital exists only if it “fits” the firm, ., is consistent with the firm’s choices of strategic stand and governance structure. Some strategists (., Hamel & Prahalad, 1994; Wright et al., 2001) point out that competitive advantage comes from aligning skills, knowledge, abilities and motives with organizational systems, structures, and processes that achieve capabilities at the organizational level (., core competence). Previous discussion of fit in SHRM literature centered around the concepts of “external fit” and “internal fit” (Becker & Gerhart, 1996; Huselid, 1995; Baird & Meshoulam, 1988;Youndt, Snell, Dean, & Lepak, 1996). External fit refers to the alignment of HR practices with external environmental demands such as the institution and the competition. Internal fit refers to the compatibility and complementarity among individual HR practices. Extending from this discussion, two issues concerning “fit” in organizational HC are explicated: 1) alignment of HC to external corporate demands and 2) compatibility of HC to internal governance structure and culture. Judge & Ferris (1992) emphasizing the importance of fit in human resources staffing decisions, argued that by selecting individuals consistent with overall business strategies, organizational performance may be enhanced. They suggested that fit can be considered in terms of a match between the competencies, goals and values of employees and those required by the organization. This notion is supported by several writers in the strategy literature (., Gupta, 1984; Hambrick & Mason, 1984) who posited that the match between the characteristics of employees and the strategic characteristics of the organization is 3 7
important in determining organizational success. When a firm’s human capital fits the organization, it produces a synergistic effect which shortens the firm’s learning curve in the market place, and improves implementation of firm strategy through single-mindedness in the corporation (Porter, 1996), giving the firm a competitive advantage over those firms still struggling to shape up their employees. Consistency among strategy, firm structure and human capital also ensures that the competitive advantages cumulate and do not erode or cancel themselves out (Porter, 1996). Therefore, “HC-organization fit” is proposed as one of the definitive dimensions of organizational human capital. Complementarity among HC. Attaining and maintaining superiority in a competitive market requires that the core employees possess a broad spectrum of complementary skills which may not reside in a single person. Thus, a combination of capabilities is required in order to successfully operate a complex organization (Barney, 1991; Carmeli & Tishler, 2004). Milgrom and his colleagues (1990, 1991, 1995) were the first to explore the economics of systems of complementary functions and activities, making observations of the tendency of companies to react to external changes with coherent bundles of internal responses, and of the need for central coordination to align functional managers. Porter (1996), while using fit and complementarity interchangeably, suggested three types of fit (., complementarity): 1) simple consistency between each activity (function) and the overall strategy (this is described in the above section relating to “fit”), 2) when activities are reinforcing, 3) optimization of effort. Similar to Milgrom and his colleagues on the notion of coordination, Porter exerts that the most basic types of effort optimization are coordination and information exchange across activities to eliminate redundancy and minimize wasted effort. Complementarity between activities obtains if “doing more of one thing increases the returns to doing (more of) the others” (Milgrom & Roberts, 1995, ), or as Porter (1996) calls it when “the whole matters more than any individual part” (p. 73). Applying the concept of complementarity to organizational level human capital, the 3 8
combined knowledge, skills, abilities, attitude and health of a firm’s workforce must form coherent systems in which individuals are mutually reinforcing, and efforts are maximized through coordination and information exchange. This requires that the organizations develop bundles of human capital with complementary expertise, skilled at dividing and reconfiguring labor among themselves, and motivated to collaborate with others. Laursen & Foss (2003) believe that complementarity is an important source of path-dependence because successful change has to involve many, if not all, relevant variables of a system and involve them in specific ways. This notion is supported by many SHRM scholars who have posited that when employees possess skills that complement each other in delivering superior services or products, it creates an interaction effect that is almost impossible to imitate by competitors (Barney & Wright, 1998; Bartlett & Ghoshal, 2002; Lado & Wilson, 1994; Pfeffer, 1994; Wright & McMahan, 1992). Therefore, “complementarity” is suggested as an important sub-construct of organizational level human capital. Specificity of HC. The degree to which human capital in a firm is specific has great implications on a firm’s internal governance structure as well as its efficiency and performance. Williamson (1981) defined human asset specificity as those that arises from “learning by doing” (p. 555). He further used whether a skill is “imperfectly transferable” to other firms as an important criterion to distinguish firm-specific human asset. He reasoned from the transaction cost proposition and argued that continuity of the employment relation in firm-specific human asset is a source of added value and need to be protected by internal governance structure (Williamson, 1981). The human capital theorists also hypothesize that specific human capital may lead to better performance through a lower turnover rate (Becker, 1964). Firm-specific human capital can be a source of sustainable competitive advantage because it takes time to develop, is immobile, and is difficult to find substitution in the market. Firm-specific human capital also constitutes a rare resource as Barney suggested in his RBV framework. Peteraf & Barney (2003) argued when a resource “is firm-specific in 3 9
nature, so that it may not be utilized readily in other settings, it may remain scarce” (p. 318). Previous literature in human capital measurement also recognized the importance of human capital specificity, as evidenced by the frequent use of these indicators such as the amount of on-job-training, job rotation and tenure (., Hitt et al., 2001; Penning et al., 1998) to gauge the level of specificity in a firm’s human capital. Summary In this chapter, the field of strategic human resource management in the context of the resourced-based view was first reviewed to shed lights on critical issues that remain in the research of HR-performance link. This review uncovered the need for a direct test of RBV hypotheses in SHRM research and a more complex view of the relationship between HR and performance, which calls for more insight on the intricate design of HR systems and their effects on possible intermediate links between HR and performance (., the black box). Human capital was proposed as the black box between HR and performance as human capital investment theory implies. The concept of human capital was then subject to an extensive review which included its origin, research applications, definitions and measurements. This review revealed several problems in the research of human capital at organizational level. To correct these problems, the literature for critical aspects of human capital that may serve as proof of its potential as the source of sustained competitive advantage for a firm was examined. Four new sub-constructs were extracted from the literature to form a new paradigm in the measurement of organizational level human capital. In the next chapter, a theoretical framework which integrates these propositions into a model of HR ROI (Return on Investment) is presented. This theoretical framework attempts to explicate the relationship among HR investment systems, organizational human capital and firm performance. 4 0
Chapter III: Theoretical Framework & Hypotheses In this chapter, a HR ROI model and hypotheses concerning how organizations improve human capital and firm performance through investment in HR practices are developed and presented. Two SHRM theoretical approaches are used to construct the HR ROI model, namely the universalistic approach and the configurational approach. Both have been proved to be valid in SHRM studies (See for example the work of Arthur, 1994; Delery & Doty, 1996; MacDuffie, 1995). Following the universalistic approach, a positive link between a firm’s overall HR investment and its performance through the mediating effect of its overall quality of human capital is proposed. Following the configurational approach, alternative hypotheses are developed by proposing that different portfolios of HR investment will be related to firm performance through the effect of these portfolios on several mediating human capital dimensions. Figure 2 presents the overall schematic representation of the theoretical framework and hypotheses. H1H4HR H2H3Firm-level HCFirmInvestmentPerformanceH5QuantityH9AcquisitionOperationalPortfolioH6FitH10FinancialH7ComplementarityDevelopmentH11PortfolioMarketH8SpecificityH12 Figure 2. Research framework and hypotheses 4 1
Universalistic Model of HR ROI The universalistic approach assumes there is a relationship between HR practice and performance that holds for all organizations and that greater use of a particular practice will always lead to better (or worse) performance (Delery & Doty, 1996). The universalistic approach is also known as the “best practices” approach. Hypotheses 1-4 are developed from the perspective of the universalistic model (see the upper portion of Figure 2). Human resource investment, in the form of HR expenditures and extensiveness of HR practices and efforts, has a positive impact on a firm’s human capital. This positive impact is demonstrated through an improved overall quality in its stock of human capital. HR Investment and Firm Performance The HR-performance link has been studied extensively in SHRM literature. Researchers in the past decade have clearly established a positive link between HR and firm performance (Arthur, 1992, 1994; Delaney & Huselid, 1996; Delery & Doty, 1996; Ellinger et al., 2002; Huselid, 1995; MacDuffie, 1995; Welbourne & Andrews, 1996). In particular, researchers have found a positive association between better management results and the use of a common set of HR practices including selective hiring & staffing, extensive training, employee participation, developmental performance appraisal, incentive compensation, etc. These practices are called the High Performance Work Practices (Huselid, 1995) or strategic HR practices (Delery & Doty, 1996) because of their strategic importance to influence firm performance. For example, Huselid (1995) revealed a relationship between High Performance Work Systems and employee turnover, gross rate of return on assets, and Tobin’s Q. This research has become a seminal study because it demonstrated that HR practices could have a 4 2
significant impact on accounting and market based measures of performance. As HR practices takes considerable time and effort to plan, design, and execute, sufficient budget and personnel effort allocation is required to ensure success. Therefore, a firm with higher level of HR investment (as measured in the level of budget allocation, personnel effort and extensiveness in the use of strategic HR practices) is more likely to establish effective HR practices that will help sustain a competitive advantage for the firm to achieve higher productivity, financial and market performance. Accordingly, the following hypothesis is proposed: Hypothesis 1: The level of a firm’s HR investment is positively related to the firm’s overall performance. HR Investment and Human Capital Organizations do not own human capital; individuals do. To legally use the human capital of an individual, organizations must pay “rent” during the period of possession of this individual’s service. The rent is equal to a market efficient wage in economic equilibrium. Anything a firm pays above and beyond the market efficient wage can be called human capital investment (Becker, 1962; Mincer, 1993). These investments include paying above-market wages, training, and thorough screening of job candidates, etc., which human capital theorists such as Becker (1962) and Schultz (1961) posit to have a significant effect in increasing human capital. Organizations rely on their workforce to produce services and products. How to increase the productive capabilities of a firm’s workforce is thus a critical management issue. HR practices, because of their functional nature to manage people, may have the most direct influence on the composition, characteristics, and behaviors of the workforce. The labor and industrial relations literature has provided much support to the link between individual HR practices and different aspects of employee characteristics. For 4 3
example, in a review of people management practices and performance, Michie & West (2005) found clear empirical evidence in the use of training to enhance task performance via its impact upon knowledge and skill, the effect of job design on psychological and physical health of employees, and the association of leadership and job involvement with worker motivation, commitment, and stress. In addition, they found that studies across work sectors consistently support the value of team-based working in promoting a sense of control, responsibility, accountability, interdependence and cooperation among employees. Therefore, a firm with higher level of HR investment (as measured in the level of budget allocation, personnel efforts and extensiveness in the use of strategic HR practices) is more likely to accrue a higher level of human capital in a firm. Accordingly, the following hypothesis is proposed: Hypothesis 2: The level of a firm’s HR investment is positively related to the overall level of the firm’s human capital. Human Capital and Firm Performance The relationship between human capital and firm performance has been theorized by many (., Becker, 1962, 1964; Mincer, 1993; Nordhaug, 1993), and empirically tested by even more scholars (See for example Carmeli & Tishler, 2004; Hitt, Bierman, Shimizu, & Kochhar, 2001; Pennings, Lee, & Van Witteloostuijn, 1998; Skaggs & Youndt, 2004; Walker, 2001; Youndt, Subramaniam, & Snell, 2004). These empirical studies lend support to the importance of human capital in many industry segments including health care, professional service, government authorities, and publicly listed companies. For instance, Carmeli & Tishler (2004) found human capital as one of the six intangible elements that explain organizational performance in 99 Israeli local authorities. Hitt et al. (2001) studying the effect 4 4
of partners’ human capital in 93 law firms found that leveraging of human capital has a positive effect on performance. Pennings et al. (1998) discovered that human capital strongly predicted firm dissolution in 1851 Dutch accounting firms. Skaggs & Youndt (2004) in a study of strategic positioning, human capital, and performance in service organizations, found considerable support to the basic argument that human capital is vital to the production and delivery processes of service organizations. Although these studies only tested a very limited aspect of human capital, ., the skills and knowledge of employees, or their proxies, there is reason to believe that an overall composition of a firm’s human capital that goes beyond employee KSAs to include also the interaction among employee KSAs and organizational strategies, tasks, cultures, and processes could yield even better explanatory power to firm performance. Thus, a positive relationship is predicted between the overall firm-level human capital and its performance. Accordingly, the following hypothesis is proposed: Hypothesis 3: The overall firm-level human capital is positively related to the firm’s overall performance. The Mediating Effect of Human Capital In search of the “black box” between HR and firm performance, researchers have conducted empirical studies to test various concepts as potential links in the HR-firm performance relationship. These include employee skills, employee behaviors, and employee motivation as shown in a review by Wright & Gardner (2000), in addition to voluntary turnover and safety as shown in the Delery & Shaw’s review in 2001. Most of these intervening variables are aspects of a firm’s human capital. Thus, it is reasonable to investigate using human capital as a mediator between HR investment and firm performance. Two theoretical perspectives also contribute to the development of this hypothesis. Human 4 5
capital theorists such as Becker (1992, 1964, 1993) and Schultz (1961) assume certain human attributes have economic value and therefore suggest investing in education and training to increase the level of these human attributes. SHRM researchers theorizing from the resource-based perspective have identified a firm’s human capital as the source of sustained competitive advantage and advocated for better management of human resource to increase the value of human capital (see for example Barney & Wright, 1998; Baron & Dreps, 1999; Bartlett & Ghoshal, 2002). Thus, a mediation hypothesis is proposed as follows: Hypothesis 4: HR investment has a positive impact on firm performance through the mediating effect of the firm’s human capital. Configurational Model of HR ROI The configurational model of HR ROI is developed as a response to the call for a more complex view of the relationship between HR and performance (Wright et al., 2001) and the need for a direct test of the RBV’s core concepts (Barney, 2001; Wright et al., 2001). There are multiple configurations of HR, each of which is designed to affect a different workforce characteristic (Delery, 1998). The configurational approach implies a HR strategy-practice fit and the equifinality effect of internally consistent HR systems (Delery & Doty, 1996). This research uses the “buy” vs. “make” HR strategy to develop two HR configurations, each of which has been claimed to increase human capital (Huselid, 1995; Youndt, Subramaniam, & Snell, 2004). The “buy” strategy builds human capital through acquiring individuals from outside the organization who have specialized knowledge and experiences. The “make” strategy builds human capital through developing its existing workforce. 4 6
Based on these two HR strategies, two HR investment portfolios are developed representing two distinct HR configurations: an acquisition portfolio and a development portfolio. The acquisition portfolio is marked by HR investment with higher concentration on acquiring human capital from outside, which may include higher budget allocation and effort in such practices as HR planning, extensive recruiting, and selective staffing. The development portfolio represents HR investment with higher intensity on developing existing human capital inside the organization, which may include higher budget allocation and effort in such HR practices as developmental performance appraisal, formal and informal training, participation, developmental feedback, and performance-based compensation. The relationship of these two HR investment portfolios and each of the four critical dimensions of firm-level human capital—., quantity of HC, HC-organization fit, complementarity among HC, and specificity of HC—will be assessed to identify the role HR plays in developing competitive advantages of firms. These four critical dimensions are proposed as a more accurate measure of the actual level, quality, and value of a firm’s human capital. They mirror the four attributes of firm resources (., rare, valuable, imperfectly imitable, non-substitutable) that create sustainable competitive advantage, as proposed by Barney (1991). Thus, examining the relationship between these four dimensions of firm level human capital and firm performance renders a direct test of the resource-based hypotheses. See Figure 2 (the lower portion) for a schematic representation of the configurational model of human resource investment ROI, which shows the hypothesized relationship between two HR investment portfolios (acquisition, development) and four dimensions of firm level human capital (quantity, fit, complementarity, specificity), as well as the proposed effect of four dimensions of firm level human capital on three aspects of firm performance (operational, financial, market). 4 7
Acquisition Portfolio and Human Capital Effective HR planning will ensure that the organization knows exactly the kind of human capital and the amount of vacancies it needs to fill in order to deliver current and future customer demands in a timely fashion. Effective planning includes the evaluation of current human capital, identification of gaps between current human capital and future needs, and implementation of a strategy to acquire employees with needed human capital (Collins, 2000). Extensive recruiting which includes the practices of using multiple sources, building a large pool of candidates, advertising and offering premium salaries and incentives should attract more job applicants. Once firms have collected larger pools of applicants, intensive selection processes can be employed to ensure that firms do hire employees with the exact KSAs, attitudes and health that add to their overall human capital. Appropriate tests and interviews will increase the amount of information gathered about each job applicant, which will help ensure the fit between job candidates and firms’ requirement. The criterion of fit is a central component of staffing decisions. Bowen, Ledford, & Nathan (1991) posited that in addition to basing selection decisions on job analysis data, it is also important to identify the dominant values, social skills, and traits necessary to fit in the organization. This can not be done without effective planning and extensive information gathering on job applicants. Once the best candidates are identified, firms may also offer above-market wages or higher wages than competitors to ensure that potential candidates do accept the job offer (Pfeffer, 1998; Snell & Dean, 1992). Firms may even consider offering sign-on bonuses to provide an additional incentive to attract employees with particularly valuable human capital. This is in alignment with the neoclassical proposition that the acquisition of valued skills leads to greater compensation (Williamson, 1981). In conclusion, HR investment in these acquisition practices may increase a firm’s human capital through acquiring a sufficient number of employees with appropriate level of human capital and enhancing the fit between individual 4 8
human capital and organizational needs. Accordingly, the following two hypotheses are proposed: Hypothesis 5: Acquisition investment portfolio has a positive impact in building sufficient quantity of firm level HC. Hypothesis 6: Acquisition investment portfolio has a positive impact in creating HC-organization fit. Development Portfolio and Human Capital Training is the major HR practice in increasing employee capabilities and has been a traditional focus in human capital theory as a means of developing firm-specific human capital (Becker, 1964; Snell & Dean, 1992). There are two types of training firms can use to increase human capital of their existing workforce: formal training and informal training. Formal training is usually a structured form of training that is designed to target specific areas of skills or knowledge deemed necessary by the organization. Formal training may appear in the form of technical training in the company’s training center, or specialized training through university classes, conferences and seminars. Informal training may include on-the-job training which uses the technique of mentoring and coaching to transfer KSAs from experienced employees to new employees, and job rotation which provides employees with increased KSAs related to other functions within the organization. Training is not the most effective means in changing employee attitudes or health. However, some training courses may increase employee awareness of safety, stress and organizational mission whereby improve employee health and attitude toward the organization. Both formal and informal training raise employee understanding in the company’s structure, production processes, 4 9
services, products, and customer needs, which are specific to the firm. For training to be effective, developmental performance appraisal should be implemented to identify lacking employee KSAs (Huselid, 1995). Performance appraisals are likely to improve employee motivation and KSAs when the appraisals objectively measure performance, provide feedback on gaps between expectations and performance, set goals for how to improve or change behavior, and are developmentally focused (Collins, 2000). When employees are highly motivated and committed to organizational goals, they are more likely to develop a sense of “esprit de corps” which contribute to the willingness to work well with each other and covering for each other. The design of compensation and reward system has significant impact on individual performance and motivation. Researchers have argued that stock ownership, profit-sharing or gain-sharing are effective tools to increase employee motivation and commitment, and to encourage employees to contribute discretionary KSAs (Huselid, 1995; Pfeffer, 1994; Stewart, 1997). Tying compensation to group and organizational performance has been found to increase cooperation and exchange of ideas among employees (Davenport & Prusak, 1998; Huselid, 1995). Another way to increase employee motivation and cooperation is through activities that promote employee participation, such as cross-functional teams and quality circles (Huselid, 1995). A sense of “esprit de corps”, cooperation, and knowledge exchange are signs of complementarity among a firm’s human capital. As the “make” strategy is synonymous to the idea of “promote from within” which provides a more stable employment relationship between a firm and its workforce and a longer working relationship among workforce members, employees are more likely to develop methods of dividing labor, cooperating, and configuring bundles of complementary skills among themselves (Pennings, Lee, & Van Witteloostuijn, 1998). Accordingly, developmental HR investment may contribute to a higher degree of complementarity and specificity of a firm’s human capital. This leads to the following two hypotheses: 5 0
Hypothesis 7: Development investment portfolio has a positive impact in promoting complementarity among firm level HC. Hypothesis 8: Development investment portfolio has a positive impact in building specificity of firm level HC. Human Capital Dimensions and Firm Performance Each of the four dimensions of organizational level HC (quantity, fit, complementarity, and specificity) has an undeniable influence on firm performance either directly or through the changed behavior of employees. A firm’s stock of human capital is what enables an organization to act (Becker, 1964). When the level of human capital is insufficient, firms will not be able to function adequately (Michie & West, 2005). The importance of HC-organization fit has been the focus of many studies in personnel psychology. Kristof-Brown, Zimmerman, & Johnson (2005) published a meta-analysis of 172 studies of individuals’ fit at work in which they show significant empirical relationships between person-job, person-organization, person-group, and person-supervisor fit with many individual-level criteria (attitudes, performance, withdrawal behaviors, strain, tenure). Argote’s study (1989) found that normative complementarity (the amount of agreement between professional groups about the norms governing their relationships) was central to the effectiveness of hospital emergency units. In a study of professional firms, Pennings, Lee, & Van Witteloostuijn (1998) found that firms emerging from a schism are more likely than de novo firms to be dissolved because of a lack of complementary knowledge and skills among professionals. They discovered that a break-up leads to “a severe partitioning of joint experiences, and loss of the home-grown complementarity of skills among professionals” (p. 5 1
428). They believe that the unraveling of those bundles of capital, most notably firm-specific human capital, is the primary reason merger firms are more likely to fail. Based on the literature, each of the four dimensions of firm-level human capital—quantity, fit, complementarity, specificity—is hypothesized to play a unique role in contributing to firm performance. Accordingly, the following hypotheses are proposed: Hypothesis 9: A sufficient quantity of firm level HC is positively related to aspects of firm performance as measured in operational, financial and market indices. Hypothesis 10: A better fit between firm requirements and firm level HC is positively related to aspects of firm performance as measured in operational, financial and market indices. Hypothesis 11: A higher level of complementarity among firm level HC is positively related to aspects of firm performance as measured in operational, financial and market indices. Hypothesis 12: A higher level of specificity of firm level HC is positively related to aspects of firm performance as measured in operational, financial and market indices. Summary In this chapter, the theoretical framework of a HR ROI model which relates firm performance to human resource investment and firm-level human capital was described and 5 2
its related hypotheses from two SHRM theoretical approaches were derived. Rationales and logics for these hypotheses were developed from existing literature. Table 5 summarizes these hypotheses. Table 5. Research hypotheses Hypothesis Universalistic Model H1: The level of a firm’s HR investment is positively related to the firm’s overall performance. H2: The level of a firm’s HR investment is positively related to the overall level of the firm’s human capital. H3: The overall firm-level human capital is positively related to the firm’s overall performance. H4: HR investment has a positive impact on firm performance through the mediating effect of the firm’s human capital. Configurational Model H5: Acquisition investment portfolio has a positive impact in building sufficient quantity of firm level HC. H6: Acquisition investment portfolio has a positive impact in creating HC-organization fit. H7: Development investment portfolio has a positive impact in promoting complementarity among firm level HC. H8: Development investment portfolio has a positive impact in building specificity of firm level HC. H9: A sufficient quantity of firm level HC is positively related to aspects of firm performance as measured in operational, financial and market indices. H10: A better fit between firm requirements and firm level HC is positively related to aspects of firm performance as measured in operational, financial and market indices. H11: A higher level of complementarity among firm level HC is positively related to aspects of firm performance as measured in operational, financial and market indices. H12: A higher level of specificity of firm level HC is positively related to aspects of firm performance as measured in operational, financial and market indices. 5 3
Chapter IV: Methodology In this chapter, the research design and the definition and measures of the studied variables are described. The research design section includes sampling and data collection method, sample description, and the data analysis procedures to test the hypotheses. The definition and measures section includes the definition and measurement design of the study variables, and scale validation of major variables through confirmatory factor analyses. Research Design Mail survey was used as the data collection method. As new models were developed to measure HR investment as a portfolio and to assess organizational human capital from the criteria of quantity, fit, complementarity and specificity, there was an urgent need to establish reliable and valid instruments for measuring both concepts. Survey questionnaires were developed first according to the result of a thorough literature review of the related concepts, as described in detail in the “literature review” chapter above and the “variable definition and measurement” section below. A panel of six HR researchers was then asked to evaluate the questionnaire in a focus group discussion to make sure that the survey items were representative of the study variables. Disagreements were discussed and resolved through consensus. In addition, two HR managers reviewed the questionnaire for face validity before full-scale data collection efforts. Sampling and Data Collection Becker & Gerhart (1996) suggested that firm-level studies are the most generalizable 5 4
and direct tests of HR-performance relationships. This study was designed to be conducted at the firm level in knowledge-intensive industries such as professional service, research and development, financial service, and hi-tech firms. Starbuck (1992) suggests that “knowledge intensive” can be applied to those firms in which knowledge has more importance than other inputs, and human capital, as opposed to physical or financial capital, dominates. Another distinctive feature of knowledge-intensive firms is the application of “expertise” (Starbuck, 1992) to solve complex problems or to provide innovative solutions. Knowledge-intensive industries were selected for several reasons. First, the nature of knowledge-intensive industries is to compete on intangible assets, particularly human capital. Therefore, it is an ideal population to test the study hypotheses. Second, knowledge-intensive industries are heterogeneous in terms of service offerings. Some knowledge-intensive firms offer a wide variety of services to customers with diverse demands; others offer a limited set of services to customers with somewhat homogeneous needs. Heterogeneity in service offerings renders rich variations in human capital composition and investment, and thus provides a fertile ground for the proposed study. Third, knowledge-intensive firms are composed mainly of knowledge workers whose work processes and contributions to the firm are relatively similar across the industries. This provides control for between-job variation across firms. These three advantages should compensate for the loss of generalizability that resulted from confining the study to knowledge-intensive industries. Because of the difficulty in obtaining company lists of knowledge-intensive industries, particularly professional service firms, in Taiwan, initially only 548 US firms listed in a company database called “Reference USA” ( October 28, 2005) were selected as the sample of this study. These were single-location professional service firms which employed between 50 and 250 employees. These companies provided professional services on a variety of subject matters including marketing, engineering, accounting, 5 5
computer systems, environmental issues, project management, training, staffing, general management, and more. Two survey questionnaires containing the measures on HR investment, organizational level human capital and firm performance were mailed to the highest level operation executive (., owner, CEO, or president) in each of these 548 firms. The two questionnaires contained identical questions in different orders for the operation executive and the HR executive. Accompanied the questionnaires were two postage-paid, self-addressed reply envelopes and a cover letter which explained the survey request and the importance of the study. (See Appendix II for a sample of the cover letter and Appendix III for a sample of the questionnaire.) In this cover letter, the operation executive was asked to forward the HR executive questionnaire to the highest level human resource executive of the company to provide HR perspective on survey items. A US$10 certificate was offered as an incentive for early replies. Follow-up mailings were sent to non-respondents approximately one month after the first mailing. As the response rate from the above sample was non-satisfactory after three months of data collection effort, the study sample was expanded to include companies in Taiwan and the US with a more flexible range of company size. Since there was no accessible list of knowledge-intensive firms in Taiwan, snowball sampling technique was used to generate responses from appropriate companies. Questions were translated into Chinese language and survey websites were set up to collect responses from the expanded sample (Chinese survey website: English survey website: Flyers and emails containing information on this study and its survey websites were distributed first through personal network to generate leads to potential targets that matched the sample definition. Telephone calls and emails were then used to confirm appropriateness of the firm type and to solicit participation of these possible leads. The study and its survey websites were also publicized 5 6
through the website of the Accounting Research and Development Foundation in Taiwan () and through emails to executive members in the management consulting, human resource, and business policy and strategy divisions of the Academy of Management in the US. At the end, a total of 129 managers and executives responded to this study, representing 105 companies. This included 18 companies from the original list of 548. Seventy companies in the original list were either moved or closed for business and thus survey questionnaires addressed to these companies were returned by the post office. This reduced the original sample to 478, resulting in a % response rate. Independent-samples t test was performed to examine non-response bias. The result showed that non-responding firms were not significantly different from those that responded to the study in terms of size, revenue, age and credit score ratings. Therefore, although the response rate was low, the responding firms were still considered representative of the initial US sample. Additional data collection effort through snowball sampling technique resulted in 820 telephone and email invitations and the participation of 12 US companies and 75 Taiwanese companies, reflecting a % response rate. Sample Description Table 6 shows a breakdown of the sample characteristics. Of the participating 105 companies, percent (n = 30) resided in the US while percent (n = 75) were Taiwan locals. Nearly half (n = 49) of the samples belonged to the professional service industry. The remaining half was evenly distributed in sales and service, financial service, R&D, high-tech manufacturing and other miscellaneous industries. About 43 percent (n = 45) of these companies categorized themselves as single practice or single product firms; while the other 57 percent (n = 60) described themselves as diversified practice or diversified product 5 7
suppliers. More than half of the sample companies (%, n = 60) were medium to small companies with less than 200 employees. Twenty one percent (n = 22) employed between 200 and 1000, and percent (n = 23) had more than 1000 employees at the time. Sales revenue figures confirmed the breakdown of firm size with percent (n = 55) making less than 20 million, 19 percent (n = 20) making between 20 and 100 million, and percent (n = 25) making more than 100 million US dollars in sales last year. In terms of firm age, percent (n = 37) were younger firms with tenure of less than 10 years. More than half ( percent, n = 61) had tenure of more than 11 years, and percent (n = 7) more than 50 years. Table 6. Sample profile Firm Characteristics Number of firms Percent Total = 105 Firm location US 30 Taiwan 75 Industry Professional Service 49 Sales & Service 11 Financial Service 13 R&D 8 Hi-tech Manufacturing 14 Other 10 Type of firm Single product 45 Diversified product 60 Firm size 1 - 50 32 (Number of 51 - 200 28 employees) 201 - 1000 22 over 1000 23 Age of firm 1 - 10 37 (Years) 11 - 20 31 21 - 50 30 over 50 7 Sales revenue under 1 17 (US$Million) 1 - 5 21 6 - 20 17 21- 100 20 over 100 25 Data missing 5 Of these 105 companies, 23 provided multiple responses. Since all respondents were 5 8
asked to rate items on all variables, decisions were made to select appropriate data source for each variable from these 23 companies basing on the following criteria: 1). responses from the highest-level operation executives were used for firm performance measures; 2). responses from human resource executives or line managers were used for human resource investment measures; 3). multiple responses on human capital measures were aggregated based on an acceptable assessment of ICC1, and interrater agreement (Rwg). ICC1 on human capital measurement was which exceeded the cutoff levels suggested by Bliese (1998) and James (1982). ICC2 was also assessed but resulted in a lower value () than the suggested threshold of . However, overall Rwg was (item average median , minimum , maximum ) which was higher than the suggested value for aggregation (James et al., 1984). As the number of responses from sampling firms varied, Box’s M test was performed to determine if firms that only supplied one response were different in their response patterns from those that supplied multiple responses. Box’s Test of Equality of Covariance Matrix tests the null hypothesis that the observed covariance matrices of the independent variables are equal across groups. If the null hypothesis is false, ., the test is significant, it can be concluded that the groups do differ in their covariance matrices. A separate Box’s M test was performed for each major study variable, ., HR acquisition, HR development, HC quantity, HC-org. fit, complementarity of HC, specificity of HC, and firm performance. All items in each study variable were entered in the test as dependent variables, and response type (., single response, multiple responses) was entered as the fixed factor. All tests resulted in a high p value (.530 for HR acquisition, .415 for HR development, .129 for HC quantity, .174 for HC-org. fit, .068 for complementarity of HC, and .117 for firm performance) except the test on specificity of HC items (p = .004). Since Box’s M was known to be very sensitive, these mostly non-significant results suggested that responses from single-response firms and multiple-response firms in most part did not differ, which 5 9
justified pooling the responses from all 105 firms together for subsequent analyses. Data Analysis Procedures This study underwent the following procedures for data analysis: 1. scale validation, 2. hypothesis testing. Scale Validation Since the study incorporated three newly defined measures with latent constructs, scale validation procedure using factor analysis for scale development was performed before these new measures were subjected to hypothesis testing. Researchers in scale development for social sciences (., Hinkin, 1998) have suggested using exploratory factor analyses (EFA) followed by a confirmatory factor analysis (CFA) on a different set of sample for this purpose. However, since the study was limited by the number of cases in the final dataset, only one factor analysis technique could be afforded. CFA was chosen in stead of the more common EFA because the new measures were developed from extensive literature research and thus were theoretically sound. A data-driven approach such as exploratory factor analysis might compromise the proposed constructs of these theory-based measures. CFA was performed using LISREL . Measurement models were constructed based on variable definition and measurement described in the previous section. Subsequently, these models were modified based on item correlation coefficients, parameter estimates, and modification indices. Items that did not provide significant contribution to the proposed constructs were dropped until satisfactory models were achieved. Several fit indices were referenced to support the use of the modified models in subsequent analyses. Construct reliability was assessed by calculating composite reliability and average variance extracted. Cronbach’s alpha reliability was used to test the internal consistency of each measure. Descriptive statistics and Pearson correlations were then used to examine the data profile and 6 0
correlations among the dimensions of HR investment, organizational level human capital, and firm performance. Hypothesis Testing Correlations of key study variables were first examined; then study hypotheses were tested using hierarchical regression analyses on SPSS. Hierarchical regression was used because this procedure allowed the researchers to control for other variables that might have an impact on the dependent variables but were not included in the research framework. Since the final sample was heterogeneous in terms of location, industry and respondent positions, these were entered in the test models as control variables in addition to company size, age, and firm type. Hypotheses 1-4 proposed relationships between overall HR investment, overall organizational level human capital and overall firm performance. Hypothesis 1 predicted a positive relationship between a firm’s HR investment and the firm’s overall performance. Three regression equations using each of the three performance measures (productivity, financial, market) as the dependent variable were formed to test this hypothesis. Each performance measure was regressed on a set of control variables and total HR investment. Hypothesis 2 proposed a positive relationship between a firm’s HR investment and its overall human capital. To test this hypothesis, overall human capital was regressed on the set of control variables and total HR investment. Hypothesis 3 stated that a firm’s human capital is positively related to its overall performance. Each of the three performance measures was regressed on the set of control variables and a firm’s overall human capital. Hypothesis 4 predicted a mediating effect of human capital between HR investment and firm performance. Mediation was tested following the three-step procedure outlined by Baron & Kenny (1986). The first step was to prove that significant relationships between HR investment and firm performance existed. Next, significant relationship between HR investment and human capital was required. Finally, in the presence of human capital, HR 6 1
investment had to show a non-significant relationship to firm performance. In hierarchical regression, this was done by entering both HR investment and human capital in the same regression model. This hypothesis would be supported if the presence of human capital reduced the β of HR investment to a non-significant level. Hypotheses 5-12 examined a finer relationship between two HR investment portfolios, four dimensions of firm level human capital, and three aspects of firm performance. Hypothesis 5-6 predicted a positive relationship between acquisition investment portfolio and the quantity and fit dimensions of firm level HC. Hypothesis 7-8 predicted a positive relationship between development investment portfolio and the complementarity and specificity dimensions of firm level HC. Each of the four dimensions of firm level HC was regressed on the set of control variables and acquisition portfolio or development portfolio respectively. Hypotheses 9-12 predicted a positive relationship between each of the four dimensions (quantity, fit, complementarity, specificity) of firm level HC and three aspects (productivity, financial, market) of firm performance. To test these hypotheses, each of the three performance measures was regressed on the set of control variables, quantity, fit, complementarity, and specificity of human capital. Variable Definition and Measurement HR Investment HR investment is the resource allocation of a firm that is directed at the firm’s human resources. HR investment includes monetary input and effort. Monetary input can be measured by a firm’s HR budget or the total spending on HR. Effort can be measured by the extensiveness of a firm’s HR practices and personnel time allocation. Two HR investment portfolios were developed from the “buy” strategy vs. “make” strategy respectively. Youndt et al. (1996) suggested that the HR practices that make up a system have an additive effect 6 2
and that firms can improve performance by increasing the number of practices or by using the practices in the system in a more comprehensive way. Therefore, the two HR investment portfolios were operationalized as additive indexes following the procedures outlined by Delery & Doty (1996) and MacDuffie (1995). Mean scores of each investment portfolio was used in the analysis. Overall HR investment was measured as the mean of combined investments in these two portfolios. Acquisition Portfolio. The acquisition portfolio was measured as the mean of monetary allocation and personnel effort toward employee acquisition activities, as well as the extensiveness in implementing those HR practices that were theoretically related to acquiring higher level of human capital. Monetary allocation in acquisition portfolio was measured as the mean of two 5-point Likert-type items: “We allot a higher budget to recruiting and staffing activities than that of the competitors,” and “We allot a higher budget to compensation and incentives than that of the competitors.” Personnel effort was measured with one Likert-type item on the degree of HR personnel devotion to recruiting and staffing activities as compared to competitors. Extensiveness of acquisition-oriented HR practices was measured as the mean of thirteen Likert-type items on the use of HR planning, extensive recruiting, selective staffing, and competitive compensation. For specific items please refer to Appendix I. Development Portfolio. The development portfolio was measured as the mean of monetary allocation and personnel effort toward employee developmental activities, as well as the extensiveness in implementing those HR practices that were theoretically related to developing higher level of human capital. Monetary allocation in development portfolio was measured as the mean of two 5-point Likert-type items: “We allot a higher budget to training and employee development than that of the competitors,” and “We allot a higher budget to employee participation programs (., quality circles, learning forums) than that of the competitors.” Personnel effort was measured with one Likert-type item on the degree of HR 6 3
personnel devotion to training, appraisal and employee participation activities as compared to competitors. Extensiveness of development-oriented HR practices was measured as the mean of thirteen Likert-type items on the use of structured training, job rotation, on-the-job training, participation, developmental performance appraisal and feedback, and performance-based compensation. For specific items see Appendix I. Organizational Level Human Capital Organizational level human capital was defined as “the level and the unique configurations of complementary employee attributes that fit the organizational value and goals in fulfilling customer demands in order to bring economic value to the firm.” Derived from this definition, the construct of quantity, fit, complementarity, and specificity were examined as four dimensions of organizational level human capital. Overall firm level HC was measured as the mean of the four dimensions (quantity, fit, complementarity, and specificity) of a firm’s human capital. Quantity of HC. Quantity of firm level HC was defined as the amount of a firm’s human capital. It was measured as the mean of four 5-point Likert-type items designed to gauge whether an organization has sufficient amount of manpower (Edvisson & Malone, 1997) and appropriate level of employee health (Michie & West, 2005; Nordhaug, 1993). Example items were “We have sufficient number of employees to handle customer demands,” and “Our employees have the appropriate physical strength to carry out their jobs.” HC-Organization Fit. HC-organization fit was defined as the degree of alignment between employee attributes and a set of corporate requirements such as firm strategy (Judge & Ferris, 1992; Porter, 1996), organizational culture (Kristof-Brown et al., 2005; Wright, 2001), and job (Kristof-Brown et al., 2005; Wright, 2001). It was measured as the mean of six 5-point Likert-type items. Human capital-strategy fit was measured with two Likert-type 6 4
items: “Our employee competence matches the company needs to help current customers,” and “Our employee competence matches the company needs in the direction it is headed.” Human capital-culture fit was measured with the following two items: “Our employee characteristics are consistent with our culture,” and “Our employees identify with the company’s value. Human capital-job fit was measured with the following two items: “Our employees possess the required competence to successfully carry out their jobs,” and “We have the right person for every job in the company.” Complementarity among HC. Complementarity refers to the system effect of combined human capital that is mutually reinforcing and effort maximizing (Porter, 1996). Complementarity is present when organizations have formed systems of internal response (., bundles of talents) (Milgrom & colleagues, 1990, 1991, 1995; Pennings et al., 1998), are efficient at coordination, division of labor, and cooperation (Milgrom & colleagues, 1990, 1991, 1995; Pennings et al., 1998; Porter, 1996), and have a sense of community and support (Argote, 1989; Porter, 1996). It was measured as the mean of nine 5-point Likert-type items. The notion of bundles of talents was measured with three items such as “Our employees form unique teams that worked great together to diagnose and solve problems,” and “The skills, knowledge and attitudes of our employees complement each other.” Three items were geared at gauging the degree of efficiency of coordination, labor division and cooperation. Examples included: “Our employees are skilled at re-configuring their talents to satisfy customer demands,” “Our employees know the strengths and weaknesses of their group members well enough to efficiently divide labor.” Three items were constructed to assess the degree of the sense of community and support within studied firms. These were: “There is a sense of community and coherence among our employees,” “Our employees enjoy working with each other to develop customer solutions,” and “Our employees exchange information with each other to minimize wasting efforts.” Specificity of HC. Three aspects of specificity were examined in this study. These 6 5
were the notion of uniqueness (Barney, 1991), proprietary employee KSAs (Williamson, 1981), and resource immobility/non-transferability (Williamson, 1981). Specificity was measured as the mean of six 5-point Likert-type items designed to gauge the degree of these three aspects in a firm. The notion of uniqueness was measured using the following two items: “The combined talents of our employees allow our company to offer unique services and products to customers,” and “The combined talents of our employees are rare in the industry.” Whether the firm possesses proprietary employee KSAs was assessed using the following two items: “The knowledge and skills of our employees are highly specific to our company,” and “Our employees are highly skilled in the company’s proprietary technology or methodology.” The degree of the firm’s human resource immobility was measured with these two items: “The knowledge and skills of our employees can not readily be used in another company, and “Our employees can easily take their talents to another company and be equally successful. (reverse coded)” Firm Performance Three types of performance measures were included in this study: operational, financial and market. These three performance measures were deemed important to firms and were analyzed in the configurational model as separate dependent variables. Overall firm performance was measured as the mean of these three performance measures. Operational performance was measured as the mean of two indices, productivity (measured as the mean of two 5-point Likert-type items on sales revenue per employee and sales growth as compared to competitors) and turnover (measured with one Likert-type item comparing the firm’s turnover rate to that of the competitors). Financial performance was measured by one Likert-type item comparing a firm’s net profit to that of the competitors. Market performance was operationalized as the firm’s reputation, which was measured as the mean of three 6 6
Liker-type items comparing the firm’s sales from referrals, sales from repeat business, and sales from new business to those of the competitors. Control Variables A firm’s human capital is likely to be highly correlated with firm age, size (Pennings et al., 1998) and industry. Firm size may also covary with HR system to affect performance (Becker & Gerhart, 1996; Huselid, 1995). Therefore, they were included in the data collection as control variables. The natural logarithmic transformation of number of full-time employees was used as the measure of firm size. The distribution resulting from the log transformation is more normal and thus is less likely to be affected by extreme scores. Additionally, it was suspected that complexity of service offering may influence firms’ demand on different types of human capital. For example, single practice service providers, such as legal consulting firms or accounting firms, may need employees with very good general knowledge of their practice and a lot of personal credibility, therefore, they may rely more on the “fit” criterion of human capital to find employees that satisfy the organizational demand for excellent academic credentials. On the other hand, solution-based service providers (., computer system consulting, management consulting, etc.) may require their employees to work well together as a team using company-specific methodologies to provide customer solution, thus, they may rely more on a complementing workforce with a high degree of company specificity. Therefore, respondents were asked to distinguish their company between “single practice service provider” and “solution-based service provider”. This variable was dummy coded. Scale Validation Confirmatory factor analysis procedure was used for scale validation on the measurement of HR investment, firm-level human capital and firm performance. A 6 7
measurement model was built for each of the three latent measures in LISREL utilizing maximum likelihood estimation. Initial models included all items described in the “variable definition and measurement” section except for those that had a low correlation (< .40) with all other items in the same measure. Modifications were then made to the model to increase the level of fit based on the data structure. Items that cross loaded with more than one latent variable or resulted in a low loading (standardized estimate < .40), a non-significant t-value at the 5% significance level, or low R2 on the proposed latent variable were dropped from the analysis (Hinkin, 1998). The measurement model was then re-specified and parameters re-estimated. This process was continued until an acceptable, substantive final model was achieved. HR Investment Measures The modified measurement model of HR investment is shown in Figure 3. This model includes two latent factors, acquisition portfolio and development portfolio, under the second-order latent variable HR. Six items contribute to the manifestation of the acquisition portfolio, while seven items constitute the measures of the development portfolio. Error covariance of HHR14 and HHR11, HHR18 and HHR12, as well as HHR30 and HHR29 were believed to be correlated because each pair measured similar HR functions, and thus were linked in the final model. Items and parameter estimates for HR investment measurement model are listed in Table 7. All items have a standard estimate of larger than with significant t-value. 6 8
Note: ?’s and d’s shown are standard estimates. Figure 3. Final two-factor HR investment measurement model 6 9
Table 7. HR investment measurement model parameter estimates Item Estimate Std. t-value Std. R2 error estimate Latent Factor: Acquisition Portfolio HHR11 We develop a large pool of applicants from which to choose for open positions. HHR12 We use incentives (., stock options, sign-on bonuses) to attract candidates HHR 13 We offer higher starting salaries than competitors to attract candidates. HHR 14 We use extensive interviews to select potential employees. HHR 18 We pay employees more than the market average. HHR 19 Salaries for core positions are higher than those of our competitors. Latent Factor: Development Portfolio HHR 20 We provide extensive specialized training to our employees for their jobs. HHR 21 We provide reimbursement for job-related conferences, seminars, journal subscriptions or association memberships. HHR 24 We use an official mentoring system for developing employees. HHR 26 We use cross-functional teams to develop company processes, methodologies or new products (services). HHR 27 We sponsor group activities such as quality circles, learning forums or company socials. HHR 29 Performance appraisals are used to set goals for employees’ development. HHR 30 We provide feedback to employees in regards to company expectations and employee performance. Note: Items not accompanied by a standard error or t-value were used as reference variables. Table 8 presents the goodness of fit statistics on the final model. The fit is considered better the closer the chi-square value is to the degrees of freedom (df) for a model (Hinkin, 1998). The chi-square to df ratio is for the final HR investment model which indicates a good fit between the observed and reproduced covariance matrices. Chi-square is non-significant (p = ) which is desirable. The root mean square error of approximation (RMSEA) equals which is indicative of reasonable fit according to Browne & Cudeck 7 0
(1993) and MacCallum et al. (1996). The goodness of fit (GFI) index of suggests an acceptable fit, although the standardized RMR of raises some doubts regarding the model’s fit as it does not meet the acceptable value of < . However, relative fit indices such as RFI and CFI indicate a reasonable fit of the model over the independence model, with values over the .90 threshold. To prove that the two latent factors in the final HR investment measurement model can be differentiated from each other, a one-factor model was constructed as shown in Figure 4. The goodness of fit statistics in Table 8 confirms that the two-factor model fitted better than the one factor-model in all categories of fit indices. Table 8. Goodness of fit statistics of final HR investment measurement models Two-factor model One-factor model df 62 62 Chi-Square P-value GFI RMSEA Std. RMR CFI RFI 7 1
?Note: ?’s and d’s shown are standard estimates. ?Figure 4. One-factor HR investment measurement model Human Capital Measures The modified measurement model of human capital is shown in Figure 5. This model includes four latent factors, quantity of HC, HC-organization fit, complementarity of HC and specificity of HC, under the second-order latent variable of human capital (HC). Error covariance of HEC2 and HEC1 were thought to be correlated because these two items both measured manpower. Error covariance of HEC24 and HEC17 were also believed to be correlated because the wording of both items referred to a combination of talents. Items and parameter estimates for human capital measurement model are listed in Table 9. All items have a standard estimate of larger than with significant t-value. 7 2
Note: ?’s and d’s shown are standard estimates. Figure 5. Final four-factor human capital measurement model 7 3
Table 9. Human capital measurement model parameter estimates Item Estimate Std. t-value Std. R2 error estimate Latent Factor: Quantity of HC HEC1 We have sufficient number of employees to handle customer demands. HEC2 We suffer from a shortage of manpower to give our customers the attention they deserve. HEC7 Our employees have the appropriate physical strength to carry out their jobs. Latent Factor: HC-organization fit HEC9 Our employee competence matches the company needs to help current customers. HEC10 Our employee competence matches the company needs in the direction it is headed. HEC11 Our employees possess the required competence to successfully carry out their jobs. HEC12 We have the right person for every job in the company. Latent Factor: Complementarity of HC HEC17 Our employees are complements of one another in their skills, knowledge and attitudes. HEC20 Our employees know the strengths and weaknesses of their group members well enough to efficiently divide labor. HEC21 There is a sense of community and coherence among our employees. HEC22 Our employees enjoy working with each other. HEC23 Our employees exchange information with each other to minimize wasting efforts. Latent Factor: Specificity of HC HEC24 The combined talents of our employees allow our company to offer unique services and products to customers. HEC25 The combined talents of our employees are rare in the industry. HEC26 The knowledge and skills of our employees are highly specific to our company. HEC27 Our employees are highly skilled in the company’s proprietary technology or methodology. Note: Items not accompanied by a standard error or t-value were used as reference variables. 7 4
Table 10 presents the goodness of fit statistics on the final model. The chi-square to df ratio is for the final human capital measurement model which indicates a good fit between the observed and reproduced covariance matrices. Chi-square value is significant (p = ) which is less desirable than a non-significant chi-square. The root mean square error of approximation (RMSEA) equals which is indicative of reasonable fit according to Browne & Cudeck (1993) and MacCallum et al. (1996). The value of the goodness of fit (GFI) index comes to and standardized RMR , which are short of the acceptable value of > and < respectively. However, relative fit indices such as CFI and RFI indicate a reasonable fit of the model over the independence model, with values over the threshold. Competing models were constructed as shown from Figures 6-8 to determine whether the four latent factors in the final human capital measurement model can be distinguished from each other. Table 10 shows that the final four-factor model is slightly better than all the competing models in most categories of fit indices, although the four-factor model does not improve much from the first three-factor model which combines quantity and fit items in the same factor. Table 10. Goodness of fit statistics of final human capital measurement models Four-factor Three-factor Three-factor Two-factor Two-factor One-factor model model 1 model 2 model 1 model 2 model df 98 99 99 101 101 102 Chi-Square P-value GFI RMSEA Std. RMR CFI RFI 7 5
Note: ?’s and d’s shown are standard estimates. Figure 6. Three-factor human capital measurement models 7 6
Note: ?’s and d’s shown are standard estimates. Figure 7. Two-factor human capital measurement models 7 7
Note: ?’s and d’s shown are standard estimates. Figure 8. One-factor human capital measurement model Firm Performance Measures The modified measurement model of firm performance is shown in Figure 9. This model includes three latent factors, operational performance, financial performance and market performance, under the second-order latent variable of firm performance (FP). Items and parameter estimates for the measurement model are listed in Table 11. All items have a standard estimate of larger than with significant t-value. 7 8
Note: ?’s and d’s shown are standard estimates. Figure 9. Final three-factor firm performance measurement model Table 11. Firm performance measurement model parameter estimates Item Estimate Std. t-value Std. R2 error estimate Latent Factor: Operational performance HFP1 Our average sales revenue per employee is higher than that of our competitors. HFP2 Our average percentage of revenue growth is higher than that of our competitors. Latent Factor: Financial performance HFP4 Our average net profit is higher than that of our competitors. HFP5 We do pretty well financially. Latent Factor: Market performance HFP6 We receive a larger volume of sales from referrals than our competitors. HFP7 We receive a larger volume of sales from repeat business than our competitors. Note: Items not accompanied by a standard error or t-value were used as reference variables. 7 9
Table 12 presents the goodness of fit statistics on the final model. The chi-square to df ratio is for the final firm performance measurement model which indicates a good fit between the observed and reproduced covariance matrices. Chi-square value is non-significant (p = > ) which is desirable. The root mean square error of approximation (RMSEA) equals which is indicative of mediocre fit according to Browne & Cudeck (1993) and MacCallum et al. (1996). The goodness of fit (GFI) index comes to and standardized RMR , which reflect a good fit. Relative fit indices such as CFI and RFI also indicate a good fit of the model over the independence model, with values well over the threshold. Figures 10 and 11 present competing two-factor models and a one-factor model as the basis for comparing the fit indices of the final model. The goodness of fit statistics in Table 12 indicates that the three-factor model performs much better than all competing models in every category of fit indices. This comparison shows that the three latent factors in the final performance measurement model are different from each other and are better at explaining the data structure than the two-factor or one-factor models. Table 12. Goodness of fit statistics of final firm performance measurement models Three-factor Two-factor Two-factor One-factor model model 1 model 2 model df 6 8 8 9 Chi-Square P-value GFI RMSEA Std. RMR CFI RFI 8 0
Note: ?’s and d’s shown are standard estimates. Figure 10. Two-factor firm performance measurement models 8 1
Note: ?’s and d’s shown are standard estimates. Figure 11. One-factor firm performance measurement model Construct Reliability While the reliability of the indicators can be examined by looking at the squared multiple correlations (R2) of the indicators, construct reliability can also be assessed by calculating a composite reliability value and the average variance extracted value for each latent variable (Diamantopoulos & Siguaw, 2000). Composite reliability values greater than are desirable (Bagozzi & Yi, 1988). Average variance extracted values less than indicate that measurement error accounts for a greater amount of variance in the indicators than does the underlying latent variable (Fornell & Larcker, 1981). Table 13 presents these values for the latent constructs in this study. Composite reliability values range from to . Therefore it can be concluded that as a set the indicators of each latent variable provide reliable measurement of the construct. Five of the nine constructs have an average variance extracted value of equal or larger than . Average variance extracted for acquisition portfolio, development portfolio, quantity of HC, and HC-organization fit variables are less 8 2
satisfactory with values ranging from to . Table 13. Construct reliability Latent construct Composite reliability Average variance extracted Acquisition portfolio Development portfolio Quantity of HC HC-organization fit Complementarity of HC Specificity of HC Operational performance Financial performance Market performance Summary This chapter presented the methodology of this research, which included the intended population and sampling for the study, the data collection method, the data analysis strategies, and the operationalization of the studied variables. Research design using survey questionnaire as the data collection instrument was described. Overall, 105 firms representing knowledge-intensive industries in the US and in Taiwan participated in the study. Sample profile showed that the professional service industry dominated the study sample. Scale validation using confirmatory factor analysis procedures using LISREL was described and final measurement models were identified for two latent factors of HR investment, four latent factors of human capital, and three latent factors of firm performance. These variables were then entered in hierarchical regression to test study hypotheses as described in data analysis strategies. The next chapter presents the analyses and results of hypothesis testing. 8 3
Chapter V: Analysis and Results Correlations Means, standard deviations, and correlations for each of the major variables appear in Table14. Cronbach’s α reliability scores were also listed for latent variables. Cronbach’s α values ranges from to indicating reasonable to good internal consistency of the sets of items in measuring the latent variables. The correlations showed initial support for the hypothesized relationships among HR investment portfolios, dimensions of firm-level human capital, and firm performance measures. All correlation coefficients were significant for these variables except for the relationships between acquisition portfolio and quantity of HC, between acquisition portfolio and market performance, and between quantity of HC and operational performance. Overall HR investment and human capital were also positively correlated with overall firm performance measure as hypothesized. Many of the company profile data correlated significantly with the study variables and were entered in the subsequent analysis as control variables. Two originally proposed control variables, firm age and firm type, were not significantly correlated with any of the study variables and were dropped from the subsequent analysis. 8 4
Table 14. Means, standard deviations, reliabilities, and correlations Variables Mean Sd. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1. Log # of - employees 2. Age of company .339** - 3. Type of firm .57 .50 .244* .013 - 4. Industry .431** .114 - 5. Firm location .71 .45 .289** .134 .451** - 6. Respondent .78 .195* .095 .290** .300** - position 7. Acquisition .82 .404** .069 .087 .042 .026 () portfolio 8. Development .82 .154 .033 .024 .055 .557** () portfolio 9. Quantity of HC .79 ** ** .170 .314** () 10. HC-org. fit .67 .046 ** ** ** .265** .413** .469** () 11. Complement- .75 ** * .289** .573** .422** .680** () arity of HC 12. Specificity of .86 ** ** .326** .524** .411** .687** .737** () HC 13. Operational .88 .233* .158 * .312** .298** .125 .416** .360** .404** () performance 14. Financial .88 .113 .257** .257** .213* .501** .435** .436** .714** () performance 15. Market .87 .008 * ** .104 .267** .202* .393** .431** .410** .475** .541** () performance 16. Overall HR .72 .317** .058 .063 .046 .883** .881** .274** .384** .488** .481** .345** .292** .210* (.88) investment 17. Firm level .63 ** ** ** .321** .558** .702** .850** .862** .872** .395** .478** .436** .498** (.93) human capital 18. Overall firm .74 .126 .025 ** .266** .323** .212* .516** .482** .492** .863** .888** .790** .334** .515** (.87) performance n = 105 * p < ** p < ( ): Cronbach’s α 85
Hierarchical Regression Hypotheses 1-4 Hypotheses 1-4 tested the relationships among overall HR investment, firm-level human capital and overall firm performance under the theorization of a universal model. Hypothesis 1 predicted a positive relationship between a firm’s HR investment and the firm’s performance. The data showed support for this hypothesis (see Table 15, Model 1 under Overall Firm Performance). HR investment was significantly related to firm performance (β = .31, p < .01) in the positive direction after controlling for firm size, industry, firm location and respondent position. To show the magnitude of the significant relationship, the returns on HR investment in the form of performance outcomes for firms with overall HR investment score one standard deviation over the mean were compared to average firms. The comparison showed an improvement of % on operational performance, % on financial performance, and % on market performance over the reported sample means respectively. Table 15. Results of regression analyses of universalistic model hypotheses Independent Overall Firm Performance Firm-level Variable Model 1Human Capital Model 2 Model 3 Log # of employees .12 .21* .19 Industry .00 Firm location .01 .12 .12 ** Respondent position ** * * ** Overall HR investment .31** .06 .52*** Firm-level HC .51*** .47*** R2 .22 .34 .34 .44 Adjusted R2 .18 .31 .30 .42 F *** *** *** *** n 105 105 105 105 Standardized regression coefficients are shown. * p < ** p < *** p < 8 6
Hypothesis 2 proposed a positive relationship between a firm’s HR investment and its overall human capital. The data showed strong support for this hypothesis (see Table 15, under Firm-level Human Capital). HR investment positively predicted firm-level human capital (β = .52, p < .001) after controlling for firm size, industry, firm location and respondent position. To show the magnitude of the significant relationship, the returns on HR investment in the form of firm-level human capital for firms with overall HR investment score one standard deviation over the mean were compared to average firms. The comparison showed a substantial increase of % on firm-level human capital over the sample mean. Hypothesis 3 stated that a firm’s human capital is positively related to its performance. Again, this hypothesis was strongly supported by the data. As shown in Table 15 under Model 2, firm-level human capital positively predicted firm performance (β = .51, p < .001) after controlling for firm size, industry, firm location and respondent position. Hypothesis 4 predicted a mediating effect of human capital between HR investment and firm performance. Mediation was tested following the three-step procedure outlined by Baron & Kenny (1986). Step one and step two of this procedure was substantiated by the regression results shown in Table 15 which linked overall HR investment to firm performance and firm-level human capital. Model 3 in table 15 shows the regression results after entering both HR investment and firm-level human capital in the same equation. The mediating effect of firm-level human capital was strongly substantiated. Comparing Model 1 and Model 3, the β weights of HR investment dropped from .31 to .06, taking it to a non-significant level in the mediation model. Hypothesis 4 was thus supported. Hypotheses 5-8 Hypothesis 5-6 predicted a positive relationship between acquisition investment portfolio and the quantity and fit dimensions of firm level HC. Hypothesis 7-8 predicted a 8 7
positive relationship between development investment portfolio and the complementarity and specificity dimensions of firm level HC. Each of the four dimensions of firm level HC was regressed on the set of control variables, acquisition investment, and development investment. As shown in Table 16 under Model 1 of each HC dimension, all these hypothesized relationships were supported by the data except for the relationship between acquisition portfolio and quantity of HC. Table 16. Regression results of human capital dimensions on HR investment portfolios IndependentQuantity of HC HC-org. Fit Complementarity Specificity of HC Variableof HC Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Log # of .13 .12 * employees Industry .14 .08 .13 .08 Firm location ** ** * ** * Respondent * ** ** ** * position Acquisition .11 .30** .33** .40*** portfolio Development .24* .40*** .58*** .55*** portfolio R2 .19 .23 .25 .33 .46 .23 .41 .27 Adjusted R2 .15 .20 .22 .29 .43 .19 .38 .23 F ** *** *** *** *** *** *** *** n 105 105 105 105 105 105 105 105 Standardized regression coefficients are shown. * p < ** p < *** p < Acquisition portfolio significantly predicted HC-org. fit (β = .30, p < .01), while development portfolio strongly predicted complementarity (β = .58, p < .001) and specificity of HC (β = .55, p < .001). To show the magnitude of the significant relationships, the returns on the two HR investment portfolios in the form of human capital dimensions for firms with investment scores one standard deviation over the mean were compared to average firms. The comparison showed acquisition portfolio scores one standard deviation over the mean increased HC-organization fit by % over the sample mean. More strikingly, development 8 8
portfolio scores one standard deviation over the mean increased complementarity of HC by % and specificity of HC by % respectively over the sample mean. Post hoc analysis tested alternative links of the two investment portfolios to the different dimensions of human capital. As shown in Table 16 under Model 2 of each HC dimension, as opposed to the non-significant acquisition portfolio, development portfolio significantly predicted quantity of HC (adjusted R2 = .20, β = .24, p < .05). Development portfolio was also better at explaining the variance of HC-org. fit with an adjusted R2 of .29 (compared to .22 of the acquisition model) and a strongly significant β-weight of .40 (p < .001). Acquisition portfolio was also tested in the models predicting complementarity and specificity of HC. Although these two models showed that acquisition portfolio was also significant in predicting these two dimensions of HC, they presented significantly less explaining power than the hypothesized models (adjusted R2 = .19, .23 respectively in predicting complementarity and specificity, as compared to .43 and .38 in the hypothesized models). Hypotheses 9-12 Hypotheses 9-12 predicted a positive relationship between each of the four dimensions (quantity, fit, complementarity, specificity) of firm level HC and three aspects (productivity, financial, market) of firm performance. To test these hypotheses, each of the three performance measures was regressed on the set of control variables, quantity, fit, complementarity, and specificity of human capital. Hypothesis 9 stated that quantity of HC would be positively related to firm performance measures. Model 1’s in Tables 17-19 show the regression results of the effect of HC quantity on firm performance measures. Although the relationships were in the positive direction, none of the β’s was significant, indicating that quantity of HC alone could not explain the variance in firm performance. Therefore, the data 8 9
did not support hypothesis 9. Table 17. Regression results of operational performance on dimensions of HC Independent Variable Model 1 Model 2 Model 3 Model 4 Model 5 Log # of employees .31** .29** .31** .33** .33** Industry .03 Firm location .05 .08 .08 .06 Respondent position ** * * * Quantity of HC .061 HC-org. fit .42*** .27* Complementarity of HC .35*** .05 Specificity of HC .41*** .25 R2 .15 .28 .25 .29 .33 Adjusted R2 .10 .25 .21 .25 .27 F ** *** *** *** *** n 105 105 105 105 105 Standardized regression coefficients are shown. * p < ** p < *** p < Table 18. Regression results of financial performance on dimensions of HC Independent Variable Model 1 Model 2 Model 3 Model 4 Model 5 Log # of employees .16 .16 .18 .20* .18 Industry .04 Firm location .03 .08 .06 .07 Respondent position Quantity of HC .17 HC-org. fit .53** .37** Complementarity of HC .44*** .13 Specificity of HC .45*** .14 R2 .08 .29 .23 .23 .31 Adjusted R2 .04 .25 .19 .19 .25 F *** *** *** *** n 105 105 105 105 105 Standardized regression coefficients are shown. * p < ** p < *** p < 9 0
Table 19. Regression results of market performance on dimensions of HC Independent Model 1 Model 2 Model 3 Model 4 Model 5 Variable Log # of employees .11 .10 .11 .13 .13 Industry * Firm location .02 .09 .06 .09 Respondent position ** * ** * * Quantity of HC .10 HC-org. fit .30** .03 Complementarity of .39*** .27 HC Specificity of HC .35*** .17 R2 .16 .23 .28 .26 .30 Adjusted R2 .12 .19 .24 .23 .24 F ** *** *** *** *** n 105 105 105 105 105 Standardized regression coefficients are shown. * p < ** p < *** p < Hypothesis 10 predicted that a higher level of HC-organization fit would be positively related to firm performance measures. The data showed strong support for this hypothesis (see Model 2’s in Tables 17-19). HC-organization fit was significantly related to operation performance (β = .42, p < .001), financial performance (β = .53, p < .001), and market performance (β = .30, p < .01). Thus, firms with a workforce high in company-needed competences are more likely to achieve better outcomes in operation, finance and market. Furthermore, the regression results in Model 2 explained more variance in firm performance (Adjusted R2 = .25, .25, .19 respectively for operation, financial and market performance) than the quantity of HC predictor shown in Model 1 of Tables 17-19 (Adjusted R2 = .101, .037, .120 respectively for operation, financial and market performance). Hypothesis 11 predicted that a complementary set of human capital would be positively related to performance measures. The data also showed strong support for this hypothesis (see Model 3’s of Tables 17-19). Complementarity of HC was significantly related 9 1
to operation performance (β = .35, p < .001), financial performance (β = .44, p < .001), and market performance (β = .39, p < .001). Thus, firms with employees that complement each other in their skills and work well with each other are more likely to achieve better outcomes in operation, finance and market. Model 3 also explained more variance in firm performance (Adjusted R2 = .208, .186, .244 respectively for operation, financial and market performance) than quantity of HC model (Adjusted R2 = .10, .04, .12 respectively for operation, financial and market performance), but was comparable to HC-org fit model in these values. Hypothesis 12 predicted that specificity of human capital would be positively related to firm performance measures. Model 4’s of Tables 17-19 show the regression results which provided very strong support for this hypothesis. Specificity of HC was found significantly related to operation performance (β = .41, p < .001), financial performance (β = .45, p < .001), and market performance (β = .35, p < .001). Thus, firms with employees that allow them to deliver rare or unique services are more likely to achieve better outcomes in operation, finance and market. The regression model explained more variance in firm performance (Adjusted R2 = .249, .191, .225 respectively for operation, financial and market performance) than quantity of HC model in Model 1 of Tables 17-19 (Adjusted R2 = .10, .04, .12 respectively for operation, financial and market performance), but was comparable to HC-org fit model and complementary HC model in these values. Although all but one of the four human capital dimensions tested positively with significant links to firm performance measures, supporting most of the study hypotheses, it was noticed that the three predictors contributed similar amount of variances in firm performance. A comparison of adjusted R2 values with the regression model using overall firm-level human capital as independent variable (adjusted R2 = .31 as shown in Model 3 of Table 15) revealed that although the aggregated overall HC construct did explain more variance in firm performance than the single dimension constructs, the resulting .31 adjusted R2 value was significantly smaller than the sum of adjusted R2 values of the four HC 9 2
dimensions in explaining variances of firm performance. These results seemed to imply that statistically at least three of the four HC dimensions were substitutes of each other in terms of predicting firm performance. To further examine this finding, post hoc tests were conducted by entering the four dimensions of HC in the same regression equations. Model 5’s in Tables 17-19 present the result of these tests. Although values of the model adjusted R2 showed a slight increase over each of the single dimension models (with the exception of a drop in R2 comparing to the model using complementarity of HC to predict market performance), these increases were mostly incremental rather than substantial. More importantly, most of the originally significant relationships in single dimension models now became insignificant in these overall models. As shown in Model 5 of Table 17, in the equation predicting operation performance, the only significant dimension was HC-org fit (β = .27, p < .05). The β-weight of the complementarity (β = .054, p > .05) and specificity (β = .248, p > .05) dimensions were absorbed by the HC-org fit dimension and reduced to a non-significant level. In the equation predicting financial performance as shown in Model 5 of Table 18, the only significant dimension was again HC-org fit (β = .37, p < .01). The β-weight of the complementarity (β = .13, p > .05) and specificity (β = .14, p > .05) dimensions were absorbed by the HC-org fit dimension and reduced to a non-significant level. In the equation predicting market performance as shown in Model 5 of Table 19, none of the dimensions were significant, although the β-weight of the complementarity (β = .27, p > .05) dimension approached significant level. β-weight of the HC-org. fit (β = .025, p > .05) and specificity (β = .172, p > .05) dimensions were absorbed by the complementarity dimension and reduced to a non-significant level. These results indicated a possible concern of collinearity. The collinearity statistics in Table 20 showed that the proportions of the variance not accounted for by other independent variables in the equation were low for the three HC dimensions in question (tolerance between .382 to .398, VIF between to ). This indicated that a larger 9 3
proportion of the variance in these dimensions could be explained by each other, and thus adding more dimensions to the equation did not contribute substantially more information to the model. An examination of correlations coefficients among HC-org fit, complementarity of HC, and specificity of HC in Table 14 also showed high correlations. Table 20. Collinearity statistics of regression models Operational Financial Market performance performance performance Tolerance VIF Tolerance VIF Tolerance VIF Quantity of HC .688 .688 .688 HC-org. fit .398 .398 .398 Complementarity of HC .382 .382 .382 Specificity of HC .383 .383 .383 Summary This chapter presented the results of data analysis. Correlations among variables of interest provided initial support for hypothesized relationships. Hypothesis testing using hierarchical regression procedures was used to control for other factors that might impact the dependent variables. Regression results were presented which showed support for most of the study hypotheses except hypothesis 5 and 9, both relating to quantity of HC. A possible collinearity concern among three human capital dimensions was also discussed. Table 21 summarizes test results of the 12 study hypotheses. 9 4
Table 21. Results of hypothesis testing Hypothesis Results H1: The level of a firm’s HR investment is positively related to the firm’s overall performance. Supported H2: The level of a firm’s HR investment is positively related to Supported the overall level of the firm’s human capital. H3: The overall firm-level human capital is positively related to Supported the firm’s overall performance. H4: HR investment has a positive impact on firm performance Supported through the mediating effect of the firm’s human capital. H5: Acquisition investment portfolio has a positive impact in Not supported building sufficient quantity of firm level HC. H6: Acquisition investment portfolio has a positive impact in Supported creating HC-organization fit. H7: Development investment portfolio has a positive impact in Supported promoting complementarity among firm level HC. H8: Development investment portfolio has a positive impact in Supported building specificity of firm level HC. H9: A sufficient quantity of firm level HC is positively related to Not supported aspects of firm performance as measured in operational, financial and market indices. H10: A better fit between firm requirements and firm level HC is Supported positively related to aspects of firm performance as measured in operational, financial and market indices. H11: A higher level of complementarity among firm level HC is Supported positively related to aspects of firm performance as measured in operational, financial and market indices. H12: A higher level of specificity of firm level HC is positively Supported related to aspects of firm performance as measured in operational, financial and market indices. 9 5
Chapter VI: Discussions and Conclusions In this chapter, data analysis results are interpreted and discussed which lead to the presentation of theoretical, research and practical implications. Contribution of the study is evaluated and described, followed by study limitations, suggested future research and conclusions. Discussions Sample and Scale This study collected data from knowledge-intensive firms which included professional service, sales and service, financial service, R&D, hi-tech manufacturing and other industry segments in the US and in Taiwan. Professional service industry was the most dominant industry segment (%) in this study. Although heterogeneous in industry, the majority of responding firms depended on human capital as their primary means to deliver products or services, and satisfied the definition of knowledge-intensive firms. The analysis and results of this research is thus germane to the knowledge-intensive industries. Statistical control for between-industry variance was used in hypothesis testing. The three study variables used new scales for measurement designed for this study. Extensive scale development process was thus undertaken to ensure the validity and reliability of the new scales. New scales were tested and modified through the procedure of confirmatory factory analysis which provided a more stringent test of validity and reliability of the scale and was better at preserving theory-based latent construct dimensions. After several modifications and re-specifications, CFA resulted in a HR investment measure with two latent factors representing acquisition portfolio and development portfolio, a firm-level 9 6
human capital measure with four latent factors representing quantity of HC, HC-organization fit, complementarity of HC and specificity of HC, and finally a firm performance measure with three latent factors representing operational, financial, and market performances. All three measures had achieved at least a mediocre fit with the current dataset as suggested by multiple fit indices. The performance measure had the best fit with current data as indicated by a high GFI of and all other fit indices within acceptable range. The human capital measure also fitted quite well with the data with RMSEA of , however, the GFI was only . This was to be expected with the higher correlations among human capital dimensions. The HR investment measure achieved reasonable fit with RMSEA of and GFI of . However, more than half of the questionnaire items in the original measure of HR investment were trimmed through the modification process because of low factor loading or cross loading, leading the researcher to suspect that maybe HR investment was better operationalized as an aggregate model rather than a latent model. This issue is further elaborated in the section of future research suggestions. The Universalistic Model In this research, relationships were found between (1) HR investment and firm performance, (2) HR investment and firm-level human capital, (3) Firm-level human capital and firm performance, and (4) HR investment and firm performance through the mediating effect of human capital. These findings are similar to the findings of other recent studies examining these relationships and confirmed the study’s theorization of a universalistic approach. The universalistic approach assumes that there is a relationship between HR practice and performance that holds for all organizations and that greater use of particular HR practices will always lead to better (or worse) performance (Delery & Doty, 1996). The universalistic approach is also known as the “best practices” approach. In this study, the final 9 7
measures for HR investment included practices of extensive recruiting, selective staffing, above-market compensation in the acquisition portfolio, and training, coaching, participation, developmental performance appraisal and feedback in the development portfolio. Inferring from the result of this study, a more extensive use of these practices in a knowledge-intensive firm should consistently improve firm performance through increased level of human capital. The Configurational Model The configurational approach implies a HR strategy-practice fit. Firms that adopt a “buy” strategy are more likely to invest in recruiting and staffing activities which make up the acquisition portfolio; whereas firms that adopt a “buy” strategy are more likely to invest in training and development activities which compose the development portfolio. The configurational model in this research took a more complex approach to examine the relationships between these two HR investment portfolios and firm level outcomes such as collective human capital and firm performance. All the hypothesized relationships in the configurational model were also supported by the current dataset except the relationships between (1) acquisition investment portfolio and quantity of firm level human capital, and (2) quantity of firm level human capital and firm performance as measured in operational, financial and market indices. Therefore, the configurational model was only partially validated. The theoretical reasoning behind the hypothesized relationship between acquisition and quantity of human capital was that extensive recruiting, which included the practices of using multiple sources, building a large pool of candidates, advertising and offering premium salaries and incentives, should attract more job applicants allowing firms to fill vacancies quicker than developing and promoting from within. However, since data were collected in the same point in time, firms which happened to have insufficient amount of manpower 9 8
might have undergone a series of recruiting events to fill vacancies, and vice versa for firms with sufficient manpower. Therefore, it was possible that when firms rated themselves as low in quantity of human capital would report a higher score in acquisition effort at the same time. Becker (1964), from a human capital theoretical perspective, concluded that a firm’s stock of human capital is what enables an organization to act. Michie & West (2005) also emphasized that when the level of human capital is insufficient, firms will not be able to function adequately. Contrary to these theorizations, relationship between quantity of firm level human capital and firm performance was not found in this sample. Two reasons offer explanation to this result. First, quantity of HC might be seen more as a hygiene factor—its importance was not recognized until it was gone. Therefore, it alone was not a good predictor of performance. Second, there might be a difference in the definition of the construct. The definition of quantity of HC was purely quantitative and was operationalized in this study as sufficient manpower to serve the customer. When Becker (1964) and Michie & West (2005) referred to “stock of human capital” and “level of human capital”, they had used the term in an overall sense and therefore might have included a quality connotation. As the quality of HC was measured in other dimensions of HC in this study, and these dimensions were proved to relate to performance in a significant way, the findings were considered consistent with the existing literature. Although not hypothesized in the study, post hoc analysis had found a significant power of development portfolio in predicting all dimensions of resource-based human capital. This seemed to lend great support to the “make” strategy of HR and was consistent with the resource-based perspective. Certainly, if a resource can be acquired from the market, it is unlikely that the resource would satisfy the resource-based criteria of being rare, inimitable or non-substitutable. Therefore, the development of a resource-based human capital measurement pretty much minimized the influence of acquisition strategy to human capital 9 9
when compared to development strategy. This finding thus should not be taken out of context and should not be used to justify improper or under-investment in HR acquisition practices. Implications Research and Theoretical Implications Universalistic model vs. configurational model. Consistent with the research of Delery & Doty (1996), both the universalistic and the configurational models were proven useful in framing HR strategies and HR practices for superior competitive advantage. The universalistic model in this research showed that higher overall HR investment had significant positive impact on overall firm-level human capital and firm performance, and is relevant at the policy level of a knowledge-intensive firm as input to budget and resource allocation decisions. The configurational model weighed the effect of an acquisition focus against that of a development focus in HR practices and showed that the development focus performed better at predicting all dimensions of human capital. The configurational findings are relevant at practice level of a knowledge-intensive firm as input to structuring HR practices for the maximum effect. Therefore, the universalistic and the configurational models in this research should not be seen as competing models but rather should be considered complements of each other in HR decisions. Potential causal links among HC dimensions. Data showed the measures of three of the four dimensions of human capital were highly correlated. It is difficult to tell at this point whether the measures were correlated because of a causal relationship, lacked discriminant validity, or plagued by respondents’ “consistency motif” (raters’ “propensity to try to maintain consistency in their responses to questions,” Podsakoff et al., 2003, p. 882). Two reverse-coded question items were designed to detect consistency motif: “We suffer 1 00
from a shortage of manpower to give our customers the attention they deserve,” and “Our employees can easily take their talents to another company and be equally successful.” The second item was dropped from the final scale because it was over-loaded with other meanings and did not do well in measuring immobility of human capital. The first item fitted quite well with other items measuring sufficient manpower and remained in the scale. The fact that most respondents correctly rated the first item suggests that consistency motif might have been less of a concern with this highly educated sample. The CFA procedure that was used to develop the final scales provided basic proof of discriminant validity. Therefore, a causal relationship among the three dimensions of human capital should be considered. It is quite logical that employees whose competencies match the organization’s need are more likely to remain in the organization longer to develop better coordination and identity with the team. Thus, HC-organization fit may lead to a higher level of complementarity in human capital. It is also possible that since complementarity takes time to develop, a complementary set of human capital is likely to be considered rare, unique and specific to the company. Therefore, complementarity of human capital may be related to a higher level of human capital specificity. Intermediate links between HC and firm performance. One of the objectives of this study was to unravel the black box between HR and firm performance. Although firm-level human capital was proved to be a strong mediator in the strategic link, there are still pieces missing from this puzzle. The overall firm-level human capital alone (controlling other profile variables) explained only .31 of the variance in firm performance, suggesting that there might be other more significant influences in the human capital and performance link. The research on the resource-based view of human asset should examine other outcome variables that may be more directly linked to human capital. Examples of intermediate links may include innovation, speed to respond to the market, customer satisfaction etc. as suggested by previous theoretical and empirical research to be related to firm performance. 1 01
Theory and research on intermediate links will help the management research field get to the bottom of the black box, and strengthen our understanding of the mechanism that transform HR efforts into positive employee behaviors and advantageous firm capabilities to succeed in the market. Practical Implications The results of the present study have several important implications for company owners, operational executives, and human resource managers. First, since HR investment is linked to overall firm performance, attention should be paid to improving the strategic decision making procedure for proper HR investment. Many companies adopt a post hoc approach to the budgeting of HR. That is, they look at the spending from previous years to develop budget for the coming year. A forward looking approach which takes into consideration the company’s strategic direction and expected outcomes is suggested. Companies can then calculate the appropriate amount of investment to produce the expected returns and budget accordingly. This study has found that on average, investment in HR one standard deviation over the mean increases operational performance %, financial performance %, market %, and overall human capital % over the mean of the entire sample. These increases are substantial and should be taken seriously in the budgetary process. Second, the study revealed that a developmental focus on HR investment exceeds the acquisition focus in predicting all dimensions of resource-based human capital. Therefore, it is advised that firms re-examine their HR strategy and practice configuration to incorporate more efforts in employee development. As suggested in the literature, training is the major HR practice in increasing employee capabilities and has been a traditional focus in human capital theory as a means of developing firm-specific human capital (Becker, 1964; Snell & 1 02
Dean, 1992). Managers can consider two types of training to increase human capital of their existing workforce: formal training, and informal training. Formal training may appear in the form of technical training in the company’s training center, or specialized training through university classes, conferences and seminars. Informal training may include on-the-job training which uses the technique of mentoring and coaching to transfer KSAs from experienced employees to new employees, and job rotation which provides employees with increased KSAs related to other functions within the organization. For training to be effective, developmental performance appraisal and feedback should be implemented. Performance appraisals are likely to improve employee motivation and KSAs when the appraisals objectively measure performance, provide feedback on gaps between expectations and performance, set goals for how to improve or change behavior, and are developmentally focused (Collins, 2000). Finally, although all dimensions except the quantity of resource-based human capital were identified to have positive significant relationships with firm performance, HC-organization fit seemed to have stood out as the most significant in predicting operational performance and financial performance, and complementarity of HC in predicting market performance. Managers are thus advised to take more proactive approach in fostering these two dimensions in their human capital. Extensive recruiting and selective staffing efforts are suggested to first find the right person for the job (criteria include the potential to “fit in” and to complement the team in terms of knowledge, skills, coordination capability and work attitude), followed by intensive development efforts to increase competence, identity with the company, and team skills. 1 03
Contributions Existing studies of organizational level human capital are plagued by inappropriate level of analysis, a mixed nature of measures that juxtaposing prediction-based indices with outcome-based indices, and the use of human capital proxies that may not accurately measure human capital. I proposed a new paradigm in measuring organizational level human capital from the resource-based perspective, in which I used the criteria of quantity, fit, complementarity and specificity to measure the value of a firm’s collective human capital at firm level. The proposed new paradigm took into consideration the many interaction effects of human capital in an organization which might have reinforced or jeopardized the productivity or efficiency of a firm. I further proposed using the link between this new paradigm and firm performance to directly test Barney’s resource-based hypotheses that resources that are rare, valuable, imperfectly imitable and non-substitutable can be a source of sustainable competitive advantage. The concept of measuring human capital in an organizational level using resource based criteria has never been introduced before, therefore this research has great potential in theoretical contribution. In addition to the theoretical contribution, this research also serves as a good start on the path to building a valid measure of the proposed resource-based human capital as an alternative to the more traditional mix-and-match measures. It also contributes to the literature by empirically presenting: 1) a positive relationship between HR investment and firm performance; 2) the mediating effect of a firm’s human capital between HR investment and firm performance; and 3) different HR investment portfolio and their returns. This study also contributes to the theory building of the resource-based perspective and strategic human resource management by confirming theoretical links of HR strategy and resources that are 1 04
valuable, rare, inimitable and non-substitutable, such as human capital, to firm performance. The study results provide company owners and business executives a good reference to form better budgetary decisions on company funds in order to acquire and build adequate human capital for desired performance. Limitations Given the exploratory nature of this study, several limitations of the research were present. First, a large portion of the dataset was collected from a single source for each unit of analysis (., an organization) using self report. This key informant method was commonly used by researchers in social sciences but is increasing criticized by OB/HR researchers as having common method variance (CMV) and reliability problems (Gerhart et al., 2000; Podsakoff et al., 2003). For this study, efforts were made to collect data from two sources at each company in order to avoid the CMV problems. However, due to reasons such as the high position requirement of the respondents, the selection of industry segments that included very small firms, the length of the survey questionnaire, and the distance of the researcher to the initial sampled firms, the portion of the responding firms that provided more than one set of data were very small. Only 20 percent of the final dataset were from firms with multiple data source, which might not be enough to prevent problems caused by CMV. CMV can inflate or deflate observed relationships between constructs, leading to both Type I and Type II errors (Podsakoff et al., 2003). Hartman’s one-factor test was used to detect the presence of CMV (Podsakoff et al., 2003; Peng et al., 2006). The test did reveal a general factor which accounted for the majority of the covariance among the measures, confirming the presence of CMV. With the threat of CMV, the observed causal relations in this study risked the possibility of being spurious rather than true causal relationships. Second, all the data were collected concurrently, thus, proving causality was not 1 05
possible. Because data for both the independent variables and the dependent variables for each organization were collected at the same point in time, causality may be opposite the direction predicted. For example, firms with higher performance may be more likely to investment in human resource. Additionally, firms low in human capital may wish to increase their human capital by investing more on their acquisition or development efforts. The later may explain why one of the hypothesized relationship in this study was not supported by the data. Hypothesis 5 stated that acquisition investment portfolio has a positive impact in building sufficient quantity of firm level HC. Although the hypothesized causal direction was theoretically logical, in practice, firms that lacked sufficient manpower would naturally increase their hiring and staffing effort. Thus, when data on the quantity of human capital were regressed on the acquisition portfolio data, the causal relationship was not significant. This problem can be solved by collecting longitudinal data or by using available historic data as independent variables. Third, with the limitation of a small sample size, measures for the newly developed latent constructs such as HR investment portfolios and firm-level human capital dimensions could not be fully validated. Exploratory factor analysis is usually used in the first stage of a statistical validation process of a new scale to establish appropriate dimensionality. Then a confirmatory factor analysis is conducted on a different set of sample to verify the new scale. Both EFA and CFA require a large sample size. The study’s 105 cases barely met the minimum requirement for either type of analysis and thus could not afford both procedures. Although the CFA procedure provided a more stringent test and was better at retaining the integrity of theory-based latent variables, this study suffered from not being able to take advantage of the full validation tests available. Collecting additional data for a re-test of the newly developed scales will provide stronger evidence of validity. Lastly, as most of the responses were generated from a snowball sampling technique, generalizing the results to a particular population can not be justified. Although the sample 1 06
consisted mainly of knowledge-intensive firms, caution is advised to generalize the results outside the sample firms. Future Research Suggestions Recent researches into the relationship between HR and firm performance are criticized of having construct validity issues and reliability problems (Delery & Shaw, 2001). The construct validity issues have to do with the inconsistency across studies regarding the specific HR practices included in the studies and differences in the measurement for same type of HR practice in different studies (Becker & Gerhart, 1996). These issues remained in this study since HR was examined as the predicting variable. HR investment was measured as a latent variable rather than a direct and observable construct in this study. Since there was no criterion reference for the scale, how true the latent variable represented the concept was questionable. The reason for choosing to measure HR investment as a latent construct was because the concept involved budgetary and monetary issues which many organizations regarded as confidential and non-disclosure company information. Likert-scale items which ask respondents to rate the degree of effort their company placed on a set of practices were less sensitive than questions asking the respondent companies to reveal the amount of spending on HR. Although this was a necessary evil, another question emerged as how the construct could be best modeled. There are still debates regarding whether HR should be measured as an aggregate model or a latent model. A latent model refers to a higher level construct with multiple underlying dimensions, while an aggregate model refers to an algebraically related set of constructs (dimensions) at the same level as the studied construct (Law, Wong, & Mobley, 1998). In the latent model, dimensions under a latent construct have to be correlated. Operationalization of the construct is usually performed through a factor analysis. In the 1 07
aggregate model, an algebraic composite of dimensions is usually reached by determining weights of each dimension from existing theories or relations between construct and other constructs in nomological network (Law et al., 1998). As this study took the latent model approach to operationalize HR investment scale, some dimensions (practices) were trimmed in the process because they did not correlate with other dimensions under the proposed latent construct. These practices were suggested by the literature to have theoretical or logical meanings to the concept of HR investment, and yet were traded for a better fit of the model to the current dataset. Therefore, the resulting scale was more data-driven than theory-driven at the final validation process. An aggregate model, however, seems better at preserving the original theoretical thinking and logic in the measurement scale. Having said that, it is difficult though to tell which model is more valid at predicting the outcome variables unless a true criterion can be found to evaluate both models in the same study. Future research on better refining the construct of HR investment and firm-level human capital is needed to fully understand the direction of causality between these concepts and their relations to sustaining organizational competitive advantage. As previously noted, the construct of HR investment can be operationalized under a latent or an aggregate model. Research is suggested to compare the validity of measures developed from these two models. Such research would substantially add to the field of SHRM toward resolving the issue of not having a consistent measure. Additionally, research on the resource-based view of human asset should examine other outcome variables that may be more directly linked to the influence of human capital. In other words, future research should consider looking into intermediate links between human capital and firm performance. For example, previous theoretical research has identified number of innovation and speed to respond to the market to be related to market performance. Further investigation can be extended to find out how complementarity of 1 08
human capital influences the creation of innovation and the speed to respond to customers. Finally, it is also suggested to replicate this study in order to better validate the constructs of HR investment and firm-level human capital and their measures without the threat of CMV. Researchers such as Peng, Kao, & Lin (2006) and Podsakoff (2003) had suggested several procedural guidelines to control CMV. These include obtaining measures of the predictor and criterion variables from different sources, building temporal, proximal, psychological, or methodological separation of measurement, and simply designing better questions and questionnaires. Traditional measures of human capital such as years of schooling and years of working experience can be added as study variables from the perspective of the traditional human capital theory in order to compare the validity of the newly developed resource-based human capital measures in predicting performance outcomes. Conclusions At a general level, this study achieved its intended goal of presenting human capital as the intermediate link between HR investment and firm performance outcomes. Specifically, the study provided evidence that HR investment is significantly related to higher level of human capital in a firm as measured by the quantity of human capital, human capital-organization fit, complementarity of human capital, and specificity of human capital. Further, HR investment is indirectly related to overall firm performance through firm-level human capital. New measurement scales were developed particularly for the HR investment construct and the human capital construct for the purpose of this research. This study also tested a more complex model linking two HR investment portfolios to four dimensions of human capital and firm performance outcomes. The findings showed that higher level of acquisition investment was linked to higher level of human 1 09
capital-organization fit, complementarity of human capital, and specificity of human capital. More significantly, a higher level of development investment was linked to higher levels of all four dimensions of human capital. In addition, each individual dimension of human capital except the quantity was found to positively predict firm performance outcomes. Limitations of the study were discussed along the line of sampling and methodological issues, which led to future research suggestions to replicate the study using procedural control of common method variance, to compare HR measures using latent construct model and aggregate construct model, to explore potential intermediate links between human capital and firm performance, and to contrast resource-based human capital measures with traditional human capital measures. This study advances understanding of the role of human capital in the HR-performance link. It also provides initial support for the use of resource-based perspective in the measurement of firm-level human capital, and thus furthers the theoretical advances in incorporating the resource-based perspective in HR research. 1 10
References Argote, L. 1989. Agreement about norms and work-unit effectiveness: evidence from the field. Basic and Applied Social Psychology, 10(2): 131-140. Arthur, J. B. 1992. The link between business strategy and industrial relations systems in American steel mills. Industrial and Labor Relations Review, 45: 488-506. Arthur, J. B. 1994. Effects of human resource systems on manufacturing performance and turnover. Academy of Management Journal, 37: 670-687. Bagozzi, R. P. & Yi, Y. 1988. On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16: 74-94. Baird, L., & Meshoulam, I. 1988. Managing Two Fits of Strategic Human Resource Management. Academy of Management Review, 13(1): 116-128. Barney, J., 1991. Firm Resources and Sustained Competitive Advantage. Journal of Management, 17(1): 99-120. Barney, J., 2001. Is the resource-based view a useful perspective for strategic management research? Yes. Academy of Management Review, 26: 41-56. Barney, J., & Wright, P. M. 1998. On becoming a strategic partner: The role of human resources in gaining competitive advantage. Human Resource Management, 37: 31-46. Baron J. N., & Dreps, D. M. 1999. Strategic Human Resources: frameworks for general managers. New York: John Wiley & Sons, Inc. Baron, R. M., & Kenny, D. A. 1986. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51: 1173-1182. Bartlett ., & Ghoshal S., 2002. Building competitive advantage through people. Sloan Management Review, 43(2): 34-41. 1 11
Becker, B. E., & Gerhart, B. 1996. The impact of human resource management on organizational performance: Progress and prospects. Academy of Management Journal, 39(4): 779-801. Becker, B. E., & Huselid, M. A. 1998. High performance work systems and firm performance: a synthesis of research and managerial implications. In G. E. Ferris (Ed.), Research in Personnel and Human Resources Management, 16: 53-102. Becker, B. E., Huselid, M. A., & Ulrich, D. 2001. The HR Scorecard: Linking people, strategy, and performance. Boston, MA: Harvard Business School Press. Becker, G. S. 1962. Investment in Human Capital: A Theoretical Analysis. Journal of Political Economy, 9-49 Becker, G. S. 1964. Human Capital. New York, NY: National Bureau of Economic Research. Becker, G. S. 1993. Human Capital: A Theoretical and Empirical Analysis with Special Reference to Education. Chicago: University of Chicago Press. Bliese, P. D. 1998. Group size, ICC values, and group-level correlations: a simulation. Organizational Research Methods, 1(4): 355-373. Bontis, N., & Fitz-enz, J. 2002. Intellectual capital ROI: A causal map of human capital antecedents and consequents. Journal of Intellectual Capital, 3(3): 223-247. Bowen, D. E., Ledford, G. E. Jr., & Nathan, B. E. 1991. Hiring for the organization, not the job. Academy of Management Executive, 5: 35-51. Brooking, A. 1996. Intellectual Capital: Core Asset for the Third Millennium. London: International Thomson Business Press. Browne, ., & Cudeck, R. 1993. Alternative ways of assessing model fit. In . Bollen & . Long (eds) Testing structural equation models (pp. 445-455). Newbury Park, CA: Sage. Carmeli, A., & Tishler, A. 2004. The relationship between organizational intangible elements and organizational performance. Strategic Management Journal, 25(13): 1257-1278. 1 12
Chan, D. 2000. Understanding adaptation to changes in the work environment: Integrating individual difference and learning perspectives, Research in Personnel and Human Resources Management, 18: 1-42. Chen, H. M., & Lin, K. J. 2003. The measurement of human capital and its effects on the analysis of financial statements. International Journal of Management, 20(4): 470-478. Colbert, B. 2004. The complex resource-based view: implications for theory and practice in strategic human resource management. Academy of Management Review, 29(3): 341-358. Collins, C. 2000. Strategic human resource management and knowledge-creation capability: examining the black box between HR and firm performance. Unpublished doctoral dissertation. University of Maryland, College Park. Crawford, R. D. 1991. In the Era of Human Capital. New York, NY: Harper Business. Davenport, T. H., & Prusak, L. 1998. Working Knowledge: How Organizations Manage What They Know. Cambridge, MA: Harvard Business School Press. Davenport, T. O. 1999. Human Capital: What Is It and Why People Invest It? San Francisco, CA: Jossey-Bass. Delaney, J. T., & Huselid, M. A., 1996. The impact of human resource management practices on perceptions of organizational performance, Academy of Management Journal, 39(4): 949-969. Delery, J. E. 1998. Issues of fit in strategic human resource management: Implications for research. Human Resource Management Review, 8: 289-309. Delery, J. E., & Doty, D. H., 1996. Modes of theorizing in strategic human resource management: Tests of universalistic, contingency, and configurational performance predictions, Academy of Management Journal, 39(4): 802-835. Delery, ., & Shaw, . 2001. The strategic management of people in work organizations: review, synthesis, and extension. Research in Personnel and Human Resources Management, 20: 165-197. 1 13
Diamantopoulos, A., & Siguaw, . 2000. Introducing LISREL. London: Sage Publications. Edvinsson, L., & Malone, . 1997. Intellectual Capital: Realizing your company’s true value by finding its hidden roots. Ellinger, A. D., Ellinger, A. E., Yang, B., & Howton, Shelly. W. 2002. The relationship between the learning organization concept and firms' financial performance: An empirical assessment, Human Resource Development Quarterly, 13(1): 5-21. Flamholtz, E. G. 1972. Assessing the validity of a theory of human resource value: a field study. Empirical Research in Accounting: Selected Studies, 241-266. Fornell, C. & Larcker, D. F. 1981. Structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18: 39-50. Gerhart, B., Wright, P., McMahan, G., & Snell, S. 2000. Measurement error in research on human resources and firm performance: how much error is there and how does it influence effect size estimates? Personnel Psychology, 53: 803-833. Guest, D. E. 1997. Human resource management and performance: a review and research agenda. International Journal of Human Resource Management, 8: 263-276. Gupta, A. 1984. Contingency linkages between strategy and general manager characteristics: a conceptual examination. Academy of Management Review, 9: 399-412. Hambrick, D. C., & Mason, P. D. 1984. Upper echelons: the organization as a reflection of its top managers. Academy of Management Review, 9: 193-206. Hamel, G., & Prahalad, C. K. 1994. Competing for the future. Harvard Business Review, 72(4): 122-129. Hinkin, T. R. 1998. A brief tutorial on the development of measures for use in survey questionnaires. Organizational Research Methods, 1: 104-121. Hitt, ., Bierman, L., Shimizu, K., & Kochhar, R., 2001. Direct and moderating effects of human capital on strategy and performance in professional service firms: A resource-based perspective, Academy of Management Journal, 44(1): 13-28. 1 14
Hudson, W. 1993. Intellectual Capital- How to Build it, Enhance it, Use it. New York, NY: John Wiley. Huselid, M. A. 1995. The impact of human resource management practices on turnover, productivity, and corporate financial performance. Academy of Management Journal, 38: 635-672. Ichniowski, C., Shaw, K., & Prennushi, G. 1997. The effects of human resource management practices on productivity: a study of steel finishing lines. American Economic Review, 87(3): 291-313. James, L. R. 1982. Aggregation bias in estimates of perceptual agreement. Journal of Applied Psychology, 67: 219-229. James, L. R., Demaree, R. G., & Wolf, G. 1984. Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69: 85-98. Joreskog, K. G., & Sorbom, D. 1989. LISREL 7: A guide to the program and applications. Chicago, IL: SPSS. Judge, T. A., & Ferris, G. R. 1992. The elusive criterion of fit in human resources staffing decisions. Human Resource Planning, 15(4): 47-67. Kocakulah, M., & Harris, D. 2002. Measuring human capital cost through benchmarking in health care environment, Journal of Health Care Finance, 29(2): 27-37. Kristof-Brown, A. L., Zimmerman, R. D., & Johnson, E. C. 2005. Consequences of individuals’ fit at work: a meta-analysis of person-job, person-organization, person-group, and person-supervisor fit. Personnel Psychology, 58(2): 281-342. Lacey, J. M. 1982. Human capital at professional organizations. Unpublished doctoral dissertation. University of California, Los Angeles. Lado, A. A., & Wilson, . 1994. Human resource systems and sustained competitive advantage: A competency-based perspective, Academy of Management Review, 19(4): 699-727. Laursen, K., & Foss, N. J. 2003. New human resource management practices, 1 15
complementarities and the impact on innovation performance. Cambridge Journal of Economics, 27: 243-263. Law, K. S., Wong, C. S., & Mobley, . (1998) Toward a taxonomy of multidimensional constructs. Academy of Management Review, 23: 741-755. Lengnick-Hall, C. A., & Lengnick-Hall, M. L. 1988. Strategic human resource management: A review of the literature and a proposed typology. Academy of Management Review, 13: 454-470. Lepak, D., & Snell, S. 1999. The human resource architecture: toward a theory of human capital allocation and development. Academy of Management Review, 24(1): 31-48. Lynn, B. E. 2000. Intellectual capital: unearthing hidden value by managing intellectual assets. Ivey Business Journal, 64(3): 48-52. MacCallum, ., Browne, ., & Sugawara, . 1996. Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1(2): 130-149. MacDuffie, J. P. 1995. Human resource bundles and manufacturing performance: Organizational logic and flexible production systems in the world auto industry. Industrial & Labor Relations Review, 48(2): 197-221. McMahan, G., Virick, M., & Wright, P. 1999. Alternative Theoretical Perspectives for strategic human resource management: Progress, problems, and prospects. In P. Wright, L. Dyer, J. Boudreau, and G. Milkovich (eds), Research in Personnel and Human Resources Management (Supplement), Greenwich, CT: JAI Press. Michie, S., & West, M. A. 2005. Managing people and performance: an evidence based framework applied to health service organizations. International Journal of Management Reviews, 5/6(2): 91-111. Miles, R. E., & Snow, C. C. 1978. Organizational strategy, structure and process. New York, NY: McGraw-Hill. Milgrom, P., Qian, Y., & Roberts, J. 1991. Complementarities, momentum, and evolution of modern manufacturing. American Economic Review, 81: 184-88. 1 16
Milgrom, P., & Roberts, J. 1990. The economics of modern manufacturing: technology, strategy, and organization. American Economic Review, 80: 511-528. Milgrom, P., & Roberts, J. 1995. Complementarities and fit: strategy, structure, and organizational changes in manufacturing. Journal of Accounting and Economics, 19: 179-208. Mincer, J. 1979. Human capital and earnings. Economic Dimensions of Education, Washington, .: National Academy of Education. Mincer, J. 1993. Studies in Human Capital: Collected essays of Jacob Mincer, Volume 1, Hants, England: Edward Elgar Publishing Ltd. Nordhaug, O. 1993. Human Capital in Organizations, Oslo, Norway: Scandinavian University Press. Nübler, Irmgard, 1997. Human Resources Development and Utilization in Demobilization and Reintegration Programs. Bonn International Center for Conversion, paper 7. Paauwe, J., & Boselie, P. 2003. Challenging ‘strategic HRM’ and the relevance of the institutional setting. Human Resource Management Journal, 13(3): 56-70. Paauwe, J., & Richardson, R. 1997. Introduction to the special issue: Strategic human resource management and performance. International Journal of Human Resource Management, 8: 257-262. Peng, T. K., Kao, Y. T., & Lin C. C. 2006. Common method variance in management research: Its nature, effects, detection, and remedies. Journal of Management, Taiwan. (Forthcoming). Pennings, J. M., Lee, K., & Van Witteloostuijn, A., 1998. Human capital, social capital, and firm dissolution. Academy of Management Journal, 41(4): 425-440. Peteraf, M. A., & Barney, J. B. 2003. Unraveling the resource-based tangle. Managerial and Decision Economics, 24(4): 309-323. Pfeffer, J. 1994. Competitive advantage through people. Boston, MA: Harvard Business School Press. 1 17
Pfeffer, J. 1998. The human equation: Building profits by putting people first. Boston, MA: Harvard Business School Press. Piazza-Georgi, B. 2002. The role of human and social capital in growth: extending our understanding. Cambridge Journal of Economics, 26: 461-479. Podsakoff, ., Mackenzie, ., & Podsakoff, . 2003. Common method bias in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88: 879-903. Porter, M. 1985. Competitive Advantage: Creating and Sustaining Superior Performance. New York, NY: The Free Press. Porter, M. 1996. What is strategy? Harvard Business Review, 74(6): 61-78. Priem, ., & Butler, ., 2001. Is the resource-based "view" a useful perspective for strategic management research? The Academy of Management Review, 26(1): 22-40. Roos, J., Roos, G., Edvinsson, L., & Dragonetti, N. C. 1998. Intellectual Capital: Navigating in the New Business Landscape. New York, NY: New York University Press. Rosen, S. 1987. Human capital. The New Palgrave: A Dictionary of Economics. London: MacMillan Press, pp. 681-689. Saint-Onge, H. 1996. Tacit knowledge: the key to the strategic alignment of intellectual capital. Strategy and Leadership, 24(2): 10-15. Schultz, T. 1961. Investment in Human Capital. American Economic Review, 1-17. Skaggs, B. C., & Youndt, M. 2004. Strategic positioning, human capital, and performance in service organizations: a customer interaction approach. Strategic Management Journal, 25(1): 85-99. Snell, ., & Dean, J. W. 1992. Integrated manufacturing and human resource management: A human capital perspective. Academy of Management Journal, 35: 467-504. Snell, ., Youndt, M. A., & Wright, P. M. 1996. Establishing a framework for research in strategic human resource management: Merging resource theory and organizational 1 18
learning. Personnel and Human Resource Management, 14: 61-90. Starbuck, . 1992. Learning by knowledge-intensive firms. Journal of Management Studies, 3:4, 262-275. Stewart, T. A. 1997. Intellectual Capita:l The New Wealth of Organizations. New York, NY: Bantam Doubleday Dell Publishing Group, Inc. Stewart, T. A. 2001. The Wealth of Knowledge: Intellectual Capital and the Twenty-first Century Organization. New York, NY: Doubleday, Random House, Inc. Storey, J. 1995. ‘HRM: still marching on, or marching out?’ in J. Storey (Ed.), Human Resource Management: A Critical Text, pp. 3-32. London: Routledge. Sullivan, P. H. 2000. Value-driven Intellectual Capital: How to Convert Intangible Corporate Assets into Market Value. New York, NY: John Wiley & Sons, Inc. Sveiby, K. E. 1997. The New Organizational Wealth: Managing & Measuring Knowledge-based Assets. San Francisco, CA: Berrett-Koehler Publishers, Inc. Walker, D. C., 2001. Exploring the human capital contribution to productivity, profitability, and the market evaluation of the firm. Unpublished doctoral dissertation of Webster University in St. Louis, USA. Welbourne, T. M., & Andrews, A. O., 1996. Predicting the Performance of Initial Public Offerings: Should Human Resource Management Be in the Equation? Academy of Management Journal, 39 (4): 891-919. Wernerfelt, B. 1984. A resource based view of the firm. Strategic Management Journal, 5: 171-180. Williamson, O. E. 1981. The economics of organization: The transaction cost approach. American Journal of Sociology, 87(3): 548-577. Wright, P. M., & Gardner, T. 2000. Theoretical and Empirical Challenges in Studying The HR Practice – Firm Performance Relationship. Paper presented at the special workshop “Strategic Human Resource Management”, European Institute for Advanced Studies in Management, INSEAD, Fontainebleau, France. March 30, 2000. 1 19
Wright, P. M., Dunford, B. B., & Snell, S. A. 2001. Human resources and the resource based view of the firm. Journal of Management, 27: 701-721. Wright, P. M., & McMahan, G. C. 1992. Theoretical Perspectives for strategic human resource management. Journal of Management, 18: 295-320. Wright, P., & Snell, S. 1998. Toward a unifying framework for exploring fit and flexibility in strategic human resource management. Academy of Management Review, 23: 756-772. Youndt, M. A., Snell, S. A., Dean, J. W., & Lepak, D. P. 1996. Human resource management, manufacturing strategy, and firm performance. Academy of Management Journal, 39(4): 836-866. Youndt, M. A., Subramaniam, M., & Snell, S. A. 2004. Intellectual Capital Profiles: An Examination of Investments and Returns. The Journal of Management Studies, 41(2): 335-361. 1 20
Appendix I Measurement Items HR investment-- Acquisition portfolio Definition Measurement Items (Scale of 1-5) Item adapted from: Budget 1. We allot a higher budget to recruiting and staffing activities than that of the competitors. 2. We allot a higher budget to compensation and incentives than that of the competitors. Effort - 3. Our HR personnel devote more time to personnel recruiting and staffing activities than our competitors. Effort - HR planning 4. We frequently evaluate the knowledge, skills, practices and abilities of our employees against current and future needs. 5. We are aware of the gaps in the knowledge, skills, and abilities of our human resource. 6. We have a plan in place to meet the human Collins, resource needs of the company. 2000 Extensive 7. We use multiple sources (., universities, Collins, recruiting newspapers, web site) to recruit job 2000; candidates. Huselid, 1995 8. We develop a large pool of applicants from Collins, which to choose for open positions. 2000; Huselid, 1995 9. We use incentives (., stock options, sign-on Collins, bonuses) to attract candidates 2000 10. We offer higher starting salaries than Collins, competitors to attract candidates. 2000 Selective 11. We use extensive interviews to select potential staffing employees. 12. We use multiple screening devices besides Collins, interviews to select potential employees. 2000 13. We only hire individuals with the exact skills, knowledge and abilities for the open positions. 14. We involve many employees in the selection Collins, process to insure that we hire individuals that 2000 will fit in. Competitive 15. We pay employees more than the market Collins, compensation average. 2000 16. Salaries for core positions are higher than Collins, those of our competitors. 2000 1 21
HR investment-- Development portfolio Definition Measurement Items (Scale of 1-5) Item adapted from: Budget 17. We allot a higher budget to training and employee development than that of the competitors. 18. We allot a higher budget to employee participation programs (., quality circles, learning forums) than that of the competitors. Effort - 19. Our HR personnel devote more time to personnel employee development and participation activities than our competitors. Effort - Structured 20. We provide extensive specialized training to Collins, practices training our employees for their jobs. 2000; Huselid, 1995 21. We provide reimbursement for job-related Collins, conferences, seminars, journal subscriptions or 2000; association memberships. Huselid, 1995 22. We provide reimbursement for continuing Collins, university education. 2000 On-the-job 23. We pair new employees with experienced Collins, training & employees. 2000; Job rotation Huselid, 1995 24. We use an official mentoring system for Collins, developing employees. 2000; Huselid, 1995 25. We use job rotation to expand the skills of Collins, employees. 2000 Participation 26. We use cross-functional teams to develop Huselid, company processes, methodologies or new 1995 products (services). 27. We sponsor group activities such as quality Huselid, circles, learning forums or company socials. 1995 28. We allow employees to have input to how their Collins, job is structured. 2000 Development29. Performance appraisals are used to set goals for Collins, al employees’ development. 2000; performance Huselid, appraisal and 1995 feedback 30. We provide feedback to employees in regards Huselid, to company expectations and employee 1995 performance. Performance-31. We had a profit-sharing (or gain-sharing) plan Huselid, based for core employees. 1995 compensation 32. Employee bonus or incentive plans are based Collins, primarily on group performance. 2000; 1 22
Huselid, 1995 Human capital Definition Measurement Items (Scale of 1-5) Item adapted from: Quantity of Amount of 1. We have sufficient number of employees to HC KSAs handle customer demands. (Edvisson & 2. We suffer from a shortage of manpower to Malone, 1997) give our customers the attention they deserve. (reverse coded) Level of 3. Our employees have the appropriate health physical strength to carry out their jobs. (Michie & West, 2005; 4. Our employees were free of job stress and psychological strains. Nordhaug, 1993) HC-Org. Fit Employee- 5. Our employee competence matches the strategy fit company needs to help current customers. (Judge, 1992; 6. Our employee competence matches the Porter, 1996) company needs in the direction it is headed. Employee- 7. Our employees possess the required job fit competence to successfully carry out their (Kristof-Browjobs. n et al., 2005; 8. We have the right person for every job in Wright, 2001) the company. Employee- 9. Our employee characteristics are consistent culture fit with the culture we intend to create. (Kristof-Brow10. Our employees identified with the n et al., 2005; Wright, 2001)company’s value. Complemen- Bundle of 11. Our employees form unique teams that tarity among talents worked great together to diagnose and HC (Milgrom & solve problems. colleagues, 12. Our customers look to us for solutions 1990, 1991, because we could organize a team of 1995; experts in different specialties to solve their Pennings et problems. al., 1998) 13. The skills, knowledge and attitudes of our employees complement each other. Division of 14. Our employees are skilled at re-configuring labor and their talents to satisfy customer demands. cooperation 15. Our employees do not collaborate with one Youndt et (Milgrom & another to develop business solutions. al., 2004 colleagues, (reverse coded) 1990, 1991, 16. Our employees know the strengths and 1995; weaknesses of their group members well Pennings et enough to efficiently divide labor. al., 1998; Porter, 1996) 1 23
Sense of 17. There is a sense of community and community coherence among our employees. and support 18. Our employees enjoy working with each (Argote, 1989; other. Porter, 1996) 19. Our employees exchange information with each other to minimize wasting efforts. Specificity Unique talent 20. The combined talents of our employees of HC (Barney, allow our company to offer unique services 1991) and products to customers. 21. The combined talents of our employees are rare in the industry. Proprietary 22. The knowledge and skills of our employees KSAs are highly specific to our company. (Williamson, 23. Our employees are highly skilled in the 1981) company’s proprietary technology or methodology. Imperfect 24. The knowledge and skills of our employees transferability can not readily be used in another company. (Williamson, 25. Our employees can easily take their talents 1981) to another company and be equally successful. (reverse coded) Firm Performance Definition Measurement Items Item adapted from: Operational Productivity 1. Our average sales revenue per Huselid, (sales per employee is higher than that of our 1995 employee) competitors. 2. Sales revenue from last year 3. Our average percentage of revenue growth is higher than that of our competitors. Turnover 4. Our average turnover rate is lower than Huselid, that of our competitors. 1995 5. Number of core employees left voluntarily last year Financial Profitability 6. Our average net profit is higher than that of our competitors. 7. We do pretty well financially. Market Reputation 8. We receive a larger volume of sales Lado & from referrals than our competitors. Wilson, 1994 9. We receive a larger volume of sales Lado & from repeat business than our Wilson, competitors. 1994 10. We receive a larger volume of sales from new business than our competitors. 1 24
Control variables: Definition Measurement Items Item adapted from: Firm size Number of 1. What is the total number of your Huselid, employees employees? 1995 Firm age 2. What is the age of your company? Type of Single practice 3. Which defines your company better? firm (., law, � Single practice (., accounting) accounting) vs. service provider Diversification � Diversified (., total solution) service provider 1 25
Appendix II Cover Letter to Operational Executive January 10, 2006 Company Address City, ST Zip Dear Mr./Ms. : I am writing to solicit your support in my dissertation research. My study investigates why some professional firms are more successful than others. The core area of investigation is the investment on human capital and how human capital contributes to company success. I believe that these topics are also of interest to you. Having being a management consultant for 15 years working for firms very much like yours, I care about how professional firms achieve their competitive advantage through people. In recent years, many practitioners and academics have seized on the importance of intangible and difficult to copy resources as a key source of advantage. Human capital is one of these resources and the most important asset of a professional firm. My study seeks to understand more fully the role human capital plays in securing profits for the firm and what firms should do to acquire or develop critical human capital. I am particularly interested in how this occurs in professional organizations, those who compete on human capital. The types of insights this research will provide will be extremely valuable to you. Participating in this research effort may in fact offer you insights into your own firm’s competitive advantages and disadvantages. I understand your time is valuable, and have worked diligently to reduce the effort required from your firm. I have enclosed two questionnaires, one for you and one for your HR manager. Both completed questionnaires are required to include your firm in this research. Each questionnaire takes about 20 minutes to complete. For your convenience, a postage-paid return envelope is also enclosed for each questionnaire. All results will be strictly confidential. Only overall results will be published and no company or individual will be able to be identified. In exchange for participation, you will receive detailed summary reports that may allow you to benchmark your firm against others in your industry. In addition, as a token of my appreciation for immediate response, companies that complete and return both questionnaires before February 6, 2006 (postmarked) are entitled to two $10 certificates. Your company is among a select sample so I am counting on you for the success of this study. Your support will be highly appreciated. If you have any questions on this research, please contact me at rosayeh@. Sincerely yours, C. Rosa Yeh Doctoral Candidate Institute of Human Resource Management National Sun Yat-sen University, Taiwan 1 26
Appendix III Survey Questionnaire Dear CEO/Owner/President: This is an important study on the role human resources play in securing profits for the firm. By participating, you will receive detailed summaries that may allow you to benchmark your firm. Both the HR’s and your responses are required to enter your company in this study. Please forward the HR survey to your HR Executive. Two $10 certificates are awarded to companies that complete and return both CEO and HR survey by February 6. As your company is among a select sample, your participation is critical to the success of this study and highly appreciated. The information you supply will be kept confidential. Please contact me at rosayeh@ for any questions. C. Rosa Yeh, Institute of HRM National Sun Yat-sen University, Taiwan, ROC Directions: These questions are intended to gauge how your company’s human capital pool serves your company, how your company acquires or develops human resources, and how well your company performs relative to your competitors. If you feel the statement truly reflect the situation in your company, circle the number 5. If you feel the statement does not reflect the situation in your company at all, circle 1. If your feelings are less strong, circle one of the numbers in the middle. Characteristics of your core employees:Strongly Strongly Disagree Agree 1. We have sufficient number of employees to handle customer 1 2 3 4 5 demands. 2. We suffer from a shortage of manpower to give our customers the 1 2 3 4 5 attention they deserve. 3. Our employees are highly skilled in the industry. 1 2 3 4 5 4. Our employees are widely considered the most talented in the 1 2 3 4 5 industry. 5. Our employees are highly motivated on their jobs. 1 2 3 4 5 6. Our employees are highly committed to the company’s success. 1 2 3 4 5 7. Our employees have the appropriate physical strength to carry out 1 2 3 4 5 their jobs. 8. Our employees were free of job stress and psychological strains. 1 2 3 4 5 9. Our employee competence matches the company needs to help 1 2 3 4 5 current customers. 10. Our employee competence matches the company needs in the 1 2 3 4 5 direction it is headed. 11. Our employees possess the required competence to successfully 1 2 3 4 5 carry out their jobs. 12. We have the right person for every job in the company. 1 2 3 4 5 1 27
Strongly Strongly Disagree Agree 13. Our employee characteristics are consistent with the culture we intend 1 2 3 4 5 to create. 14. Our employees identified with the company’s value. 1 2 3 4 5 15. Our employees form unique teams that work great together to 1 2 3 4 5 diagnose and solve problems. 16. Our customers look to us for solutions because we can organize a 1 2 3 4 5 team of experts in different specialties to solve their problems. 17. Our employees are complements of one another in their skills, 1 2 3 4 5 knowledge and attitudes. 18. Our employees are skilled at re-configuring their talents to satisfy 1 2 3 4 5 customer demands. 19. Our employees do not collaborate with one another to develop 1 2 3 4 5 business solutions. 20. Our employees know the strengths and weaknesses of their group 1 2 3 4 5 members well enough to efficiently divide labor. 21. There is a sense of community and coherence among our employees. 1 2 3 4 5 22. Our employees enjoy working with each other. 1 2 3 4 5 23. Our employees exchange information with each other to minimize 1 2 3 4 5 wasting efforts. 24. The combined talents of our employees allow our company to offer 1 2 3 4 5 unique services and products to customers. 25. The combined talents of our employees are rare in the industry. 1 2 3 4 5 26. The knowledge and skills of our employees are highly specific to our 1 2 3 4 5 company. 27. Our employees are highly skilled in the company’s proprietary 1 2 3 4 5 technology or methodology. 28. The knowledge and skills of our employees can not readily be used in 1 2 3 4 5 another company. 29. Our employees can easily take their talents to another company and 1 2 3 4 5 be equally successful. HR pStrongly ractices for your core employees: Strongly Disagree Agree 1. We allot a higher budget to recruiting and staffing activities than that 1 2 3 4 5 of the competitors. 2. We allot a higher budget to compensation and incentives than that of 1 2 3 4 5 the competitors. 3. We allot a higher budget to training and employee development than 1 2 3 4 5 that of the competitors. 1 28
Strongly Strongly Disagree Agree 4. We allot a higher budget to employee participation programs (., 1 2 3 4 5 quality circles, learning forums) than that of the competitors. 5. Our HR personnel devote more time to recruiting and staffing 1 2 3 4 5 activities than our competitors. 6. Our HR personnel devote more time to employee development and 1 2 3 4 5 participation activities than our competitors. 7. We frequently evaluate the knowledge, skills, and abilities of our 1 2 3 4 5 employees against current and future needs. 8. We are aware of the gaps in the knowledge, skills, and abilities of our 1 2 3 4 5 human resource. 9. We have a plan in place to meet the human resource needs of the 1 2 3 4 5 company. 10. We use multiple sources (., universities, newspapers, web site) to 1 2 3 4 5 recruit job candidates. 11. We develop a large pool of applicants from which to choose for open 1 2 3 4 5 positions. 12. We use incentives (., stock options, sign-on bonuses) to attract 1 2 3 4 5 candidates 13. We offer higher starting salaries than competitors to attract 1 2 3 4 5 candidates. 14. We use extensive interviews to select potential employees. 1 2 3 4 5 15. We use multiple screening devices besides interviews to select 1 2 3 4 5 potential employees. 16. We only hire individuals with the exact skills, knowledge and abilities 1 2 3 4 5 for the open positions. 17. We involve many employees in the selection process to insure that we 1 2 3 4 5 hire individuals that will fit in. 18. We pay employees more than the market average. 1 2 3 4 5 19. Salaries for core positions are higher than those of our competitors. 1 2 3 4 5 20. We provide extensive specialized training to our employees for their 1 2 3 4 5 jobs. 21. We provide reimbursement for job-related conferences, seminars, 1 2 3 4 5 journal subscriptions or association memberships. 22. We provide reimbursement for continuing university education. 1 2 3 4 5 23. We pair new employees with experienced employees. 1 2 3 4 5 24. We use an official mentoring system for developing employees. 1 2 3 4 5 25. We use job rotation to expand the skills of employees. 1 2 3 4 5 26. We use cross-functional teams to develop company processes, 1 2 3 4 5 methodologies or new products (services). 1 29
Strongly Strongly Disagree Agree 27. We sponsor group activities such as quality circles, learning forums or 1 2 3 4 5 company socials. 28. We allow employees to have input to how their job is structured. 1 2 3 4 5 29. Performance appraisals are used to set goals for employees’ 1 2 3 4 5 development. 30. We provide feedback to employees in regards to company 1 2 3 4 5 expectations and employee performance. 31. We have a profit-sharing (or gain-sharing) plan for core employees. 1 2 3 4 5 32. Employee bonus or incentive plans are based primarily on group 1 2 3 4 5 performance. Firm PerformanceStrongly Strongly : Disagree Agree 1. Our average sales revenue per employee is higher than that of our 1 2 3 4 5 competitors. 2. Our average percentage of revenue growth is higher than that of our 1 2 3 4 5 competitors. 3. Our average turnover rate is lower than that of our competitors. 1 2 3 4 5 4. Our average net profit is higher than that of our competitors. 1 2 3 4 5 5. We do pretty well financially. 1 2 3 4 5 6. We receive a larger volume of sales from referrals than our 1 2 3 4 5 competitors. 7. We receive a larger volume of sales from repeat business than our 1 2 3 4 5 competitors. 8. We receive a larger volume of sales from new business than our 1 2 3 4 5 competitors. Directions: The following background information will help the researchers place your company in the appropriate group for comparison. Please fill in the blanks or check the appropriate boxes. 1. What is the total number of your employees? 2. What is the age of your company? 3. Sales revenue from last year: US$ million 4. Number of core employees left voluntarily last year: 1 30
5. Which defines your company better? � Single practice (., accounting) service provider � Diversified (., total solution) service provider 6. Your title/position in the company: Thank you for completing this survey. Your participation is truly appreciated. Please make sure you have answered every question and return this survey via the enclosed postage-paid envelope at your earliest convenience. As a token of appreciation for immediate response, I will send you a $10 certificate when I have received both questionnaires from your company by February 6, 2006 (postmarked). Please fill in your name and e-mail address to receive summary results of this study and your $10 certificate. If you prefer traditional mail delivery, please fill in your postal address. Name: e-mail: Address: 1 31
