This study examines determinants of professional human resource management (HRM) practices within a sample of approximately 700 small to medium-sized firms. Predictions from the agency theory and the resource-based view of organizations lead to alternate hypotheses regarding the direct and indirect negative effects of family ownership and management on the usage of professional HRM practices. Results support predictions for both direct and indirect effects. These indirect effects occur through intermediary variables that reflect organizational complexity, such as firm size, (the presence of a) formal business plan, and HRM specialization. The findings lend partial support to both theories.
Human resource management (HRM) has been defined as the "process of attracting, developing and maintaining a talented and energetic workforce to support organizational mission, objectives, and strategies" (Schermerhorn 2001, p. 2400). Effective HRM practices are becoming increasingly important in the modern knowledge-based economy, as companies face the double challenge of the need for more highly trained employees and the shortage of qualified labor. These challenges, coupled with the third trend toward smaller firms in general, reinforce the need for effective HRM practices in the small firm (Audretsch and Thurik 2001).
Empirical research confirms that in general, smaller firms make less use of professional HRM practices than larger firms. For example, smaller firms make less use of formalized recruitment practices, provide less training to their employees, and are less likely to use formalized performance appraisals. Despite the size effect, research suggests that far from being homogeneous, small firms vary widely in the professional HRM practices in use (De Kok and Uhlaner 2001).
Variation in family ownership and management may help to explain the differences in HRM practices among small and medium-sized enterprises (SMEs). Dyer (2003) and Schulze, Lubatkin, and Dino (2003) all point out that the family is a neglected variable in organizational research. Nevertheless, a research stream is emerging that generally confirms a negative relationship between family firm governance and the use of professional HRM practices (Fiegener et al. 1996; Cyr, Johnson, and Welbourne 2000). In this paper, we pursue this research stream further by deriving and testing a model to explain whether and why family-owned and managed firms (1) tend to use fewer professional HRM practices than other SMEs. In developing the propositions of the model, we compare and contrast predictions and explanations based on agency theory with those based on the resource-based view. We argue that the direct effect of family ownership and management on the types of HRM practices found in SMEs is consistent with agency theory whereas indirect effects, via various organization characteristics variables associated with greater organizational complexity and resource richness, may support a resource-based view of the firm. The organization characteristics chosen for the present study include firm size, formal business planning, HRM specialization, and export strategy. In addition, we control for other organization characteristics such as firm age, sector, unionization, and franchising. Before presenting the methodology and results of an empirical study of approximately 700 SMEs, we will review aspects of the literature that provide conceptual support for the proposed model and hypotheses.
Professional HRM Practices
One of the lingering questions in HRM research is whether or not there is a single set of policies or practices that represents a universally superior approach to managing people. Theories on best practices or high commitment theories suggest that universally, certain HRM practices, either separately or in combination with others, are associated with improved organizational performance. They maintain that well-paid, well-motivated workers, working in an atmosphere of mutuality and trust, generate higher productivity gains and lower unit costs (Boxall 1996; Pfeffer 1994). These HRM practices have therefore been labeled as "best practices," "high performance practices," or "high commitment workplace systems." Some empirical research supports this view. For instance, Huselid (1995) reports from a sample of 968 firms that those using comprehensive employee recruitment and selection procedures, extensive employee involvement and training, and formal performance appraisal approaches linked to incentives are likely to have lower employee turnover, higher productivity, and enhanced corporate financial performance.
Others, however, argue that a contingency approach is more likely to generate superior performance, with empirical research also supporting this perspective (Cappelli and Crocker-Hefter 1996; Meyer, Tsui, and Hinings 1993; Lawler and Jenkins 1992). Research to date, though informative, does not unequivocally support either a best practices or a contingency view. Additional sampling is required to represent size, sector, and governance structures more fully. In addition, these need to be tested for the full range of HRM practices.
Because of the limitations of the research to date, we have decided to avoid using such labels as "best practice" and "high performance" for a defined set of HRM practices because doing so seems premature (especially in the context of SMEs). Using the term "formal" with reference to such HRM practices (De Kok and Uhlaner 2001; Heneman and Berkley 1999; Aldrich and Langton 1997) is also problematic. The term "formal" takes on multiple meanings in the literature, for example, whether a practice is written, standardized, and/or defined by the employer. Certain HRM practices (for example, performance pay, or referrals by employees) may be considered appropriate by experts but not necessarily formal, according to one or more of these definitions.
The term that seems to be most suitable for our study is "professional" HRM practices (Gnan and Songini 2003; Matlay 1999). The HRM practices chosen for our study are derived primarily from experts in the field of HRM, whether or not the practice has been empirically validated against performance indicators within SMEs. Such practices typically conform to legal requirements and professional standards established in a number of western economies (and listed in such standardsetting bodies as the Society for Human Resource Management in the United States). We have therefore decided to use this term for our present study.
Family Ownership and Management and Professional HRM Practices: Empirical and Theoretical Perspectives
Past research confirms a negative relationship between family ownership and management and professional HRM practices and expertise. For instance, Aldrich and Langton (1997) find a negative relationship between the number of family members who work in a firm and formal HRM practices. Fiegener et al. (1996) confirm a negative relationship for promotion decisions. Though non-family firms emphasize outside work experience and university training in promotion decisions, family firms rarely do so. Research by Reid and Adams (2001) confirms this pattern. In a study of Irish SMEs ranging in size from 20 to 100 employees, they find that family businesses are less likely to have professional HRM practices, including the use of references, appraisal systems, a peer appraisal process, training assessment, or merit-based pay.
Past researchers have drawn upon a wide variety of theories to explain the differences in professionalism of management and/or HRM practices between firms, and specifically between family and nonfamily-owned firms. In this study, we will focus on two theories in particular: the agency theory and the resource-based view.
Agency Theory and the Family Firm
Agency theory examines the relationship between principals and agents, often representing owner(s) and manager(s) of an organization. Recently, agency theory is being used to study family firms: whenever the owner and manager are part of the same family, coordination between the two (for example, through contracts and monitoring) should be more efficient and thus, less expensive (Steier 2003). If the owner and manager are one and the same, monitoring is not even necessary, saving even more on agency costs.
At a lower hierarchical level (managers versus employees), agency theory may also be useful for the study of HRM practices (Randoy and Goel 2003; Heneman, Tansky, and Camp 2000). For employees who belong to the same family as the owner and managers, the same logic as that mentioned previously suggests that less professional HRM practices are required to align the interests of managers and employees. This may also hold for employees who are not related to the owner and/or managers, to the extent that family-owned enterprises are able to create an organizational culture where all employees feel they belong to the same family (Pollak 1985).
However, it is not always the case that employees are loyal towards the family and/or perform well. In the case of a relative not performing well, other family owner/managers may be more reluctant to take action than against a nonperforming nonrelative for fear of damaging family relations, even if it is bad for the business. Schulze, Lubatkin, and Dino (2003) refer to this latter phenomenon as an altruism problem: a situation where the owner/manager, by attempting to help other family members, unintentionally and/or indirectly encourages them to shirk their duties. Nevertheless, such altruism, although leading to negative performance, does not necessarily change the impact of family firm governance on the types of monitoring devices used (Gomez-Mejia, Nunez-Nickel, and Gutierrez 2001).
Following these arguments, we suggest that family firms have less of a need to monitor agents in the firm, especially when they are from the same family. Even if their performance is appraised, the basis for rewarding family employees is less likely to be related solely on their performance, and thus makes a professional compensation system also unnecessary. At least in smaller firms, recruitment is also simplified to the extent that family members are chosen over nonfamily members. Since the family owner-manager's expectations and goals influence the choice of HRM policy, it is thus seen as less likely that (especially smaller) family firms will choose professional HRM policies.
The Family Firm and the Resource-Based View
An alternative explanation for the relationship between family ownership and management and professional HRM practices is grounded in the resource-based view of the firm. The resource-based view is based on the assumption that differences in physical, organizational, and human resources between firms cause a fundamental heterogeneity in their productive potential (Priem and Butler 2001). It is one of the main theoretical perspectives of HRM research, and is recently also used to understand the relationship between family ownership and management and other organizational characteristics (Sirmon and Hitt 2003).
Reid and Adams (2001) found that many family firms use less professional HRM practices, and explain this by suggesting that such firms have more limited organizational capabilities. In this paper, we further develop their explanation, and argue that the resource-based view can be used to suggest an indirect effect of family ownership and management on professional HRM practices. Especially among SMEs, family ownership and management may be negatively associated with professional HRM practices because of resource limitations of family firms. These limitations follow from their comparatively smaller size and reduced complexity. In our study, variables which reflect this reduced complexity (in addition to firm size) include a (less) specialized HRM department or staff, less formal planning, and lack of an export strategy. In the following discussion, we will discuss the rationale for our argument.
First of all, family firms are often smaller than nonfamily firms (Cromie, Stephenson, and Monthieth 1995; Daily and Dollinger 1993). They are also less complex than nonfamily firms in that they are less specialized, less likely to have an HRM department (Reid and Adams 2001), and less likely to use formal accounting and planning practices than nonfamily firms (2) (Jorissen et al. 2002).
Second, the resource-based view supports a relationship between firm size and organizational complexity on the one hand, and professional HRM practices on the other. The remainder of this section provides a brief overview of literature on this subject, where specialization, formal planning, and export strategy will be used as specific indicators of organizational complexity.
Firm Size and Professional HRM Practices. The link between firm size and professional HRM practices is well established (De Kok 2003). Firm size is often used as a general indicator for the lack of specific resources, for example, financial, organizational, and/or human resources. For example, most professional HRM practices require considerable development costs (Klaas, McClendon, and Gainey 2000). This results in a cost advantage for larger firms, which is strengthened by the limited supply of financial resources of many small firms. Others argue that it is the lack of specific knowledge (particularly the recognition of the importance of HRM issues) that leads to less use of professional HRM practices in small businesses (Golhar and Deshpande 1997; Hendry et al. 1991). In other words, it is not the small number of employees that explains the lack of professional HRM practices, but the lack of specific organizational and human resources.
Organizational Complexity and Professional HRM Practices. Company growth theories (Gnan and Songini 2003; Chandler and McEvoy 2000) point towards the positive correlations between firm size, complexity, and professional HRM practices. As firms increase in size and complexity, they typically develop more layers of management and more formalized and/or systematized procedures and policies in order to process information more effectively within the organization. There are various explanations for this finding. For example, larger companies have a greater demand for human resources, and (thus, we assume) therefore, a greater demand for specific HRM practices such as recruitment, selection, and performance appraisal. This stimulates more professional development of these practices.
Specialization is typically associated with greater firm size (Wagar 1998; Jackson, Schuler, and Rivero 1989). Employees in smaller firms often have to perform a greater variety of tasks than do employees in larger firms, and specialists are less likely to be found in smaller firms. For example, firm size is positively related to HRM specialization (that is, the presence of a specific HRM department and/or manager) (Cyr, Johnson, and Welbourne 2000; Heneman and Berkley 1999). Daman-pour (1996) provides an explanation for why specialization (or more generally, structural complexity) is consistent with the resource-based view. "In complex organizations, coalitions of specialists in differentiated subunits increase the depth of the knowledge base, which in turn, increases the development of new ideas" (Damanpour 1996, p. 695). In other words, greater specialization is associated with greater knowledge resources. This assumption is indirectly confirmed by research by Damanpour (1996) using a meta-analysis from over 20 published studies showing a positive relationship between specialization and innovation.
Some research using the resource-based view sees strategic planning as a type of organizational capability (Michalisin, Smith, and Kline 1997). Research specifically linking formal business planning and/or strategic planning to professional HRM practices is quite limited. However, in light of the previous discussion, it is logical to infer that firms with the organizational capabilities and resources to develop formal plans are also more likely to have the resources to develop professional HRM practices. Furthermore, the availability of a business plan can be interpreted as a characteristic of organizations with a relatively long planning horizon. These firms will be more aware of the need to use professional HRM practices to build a competent employee base, implying a relatively high perceived value of HRM practices by the chief executive officer (CEO). In addition, the presence of a formalized plan may reflect overall formalization levels in the organization.
Recent research on exporting links the existence of an export strategy with greater resource availability. The presence of an export strategy within an SME may indicate an organization's ability to handle greater complexity and environmental uncertainty, which require access to more resources. For instance, Julien and Ramangalahy (2003) conclude that SMEs' limited capacity to acquire information and use sources is a major factor in explaining their low level of involvement and performance in export markets. The exporting performance of SMEs is determined in part by their ability to acquire and manage foreign market information.
Model and Hypotheses
In this section, we present the model and hypotheses to be tested in our research. The main characteristic of this model is that it distinguishes between a direct and indirect effect of family ownership and management on professional HRM practices. (See Figure 1).
H1 predicts an indirect (negative) relationship between family ownership and management (referred to as the family firm variable in the model) and professional HRM practices via certain organization characteristics associated with greater complexity and/or richer organizational resources (arrows 1 and 2 in Figure 1). The assumptions for this hypothesis draw upon the resource-based model and research on organization complexity and uncertainty to suggest that family firms may use fewer professional HRM practices because they have fewer resources (typically being smaller) and therefore, are less complex. We state H1 as follows:
H1: Family firms apply less professional HRM practices because of differences in certain organization characteristics associated with organizational complexity an/or resource availability.
To test H1, we include a limited number of variables to represent these organization characteristics, including firm size, the presence of formal business plans, HRM specialization either by department or individual manager, and export strategy.
H2 predicts that at least part of the differences in the extent to which professional HRM practices are used by family and nonfamily firms cannot be attributed to differences in organization complexity or resource availability but rather to a direct effect of the family firm variable consistent with agency theory predictions (arrow 3 in Figure 1). We state H2 as follows:
[FIGURE 1 OMITTED]
H2: Family firms, even when controlling for certain organization characteristics associated with complexity and/or resource availability, are likely to have less professional HRM practices than similar nonfamily firms.
This section discusses the collection of the necessary data, and the variables used in the analyses.
Sample and Data Collection
Data was collected by means of a written questionnaire sent to Dutch small and medium-sized enterprises. The questionnaire was developed by the University of Southern Queensland, Australia (Wiesner and McDonald 2001). A first version of the questionnaire was submitted to a sample of 70 Australian SMEs, 31 of which responded. The results of this pilot were used to modify the questionnaire. Subsequently, it was submitted to several senior Australian academics in HRM for their comments. The revised questionnaire was then translated and further revised by Dutch HRM researchers and practitioners.
A stratified sample plan was drawn, distinguishing six sectors (manufacturing, construction, trade and repairs, catering, transport and communication, services) and three size classes (20-49 employees, 50-99 employees and 100-199 employees). Not all respondents fell within the originally defined sample classes. One hundred enterprises have either less than twenty or more than 200 employees. To avoid the loss of these observations, we decided to apply the small business administration definition of SMEs (Flanagan and Deshpande 1996), and to include all enterprises with 1-500 employees in our analysis.
Four thousand questionnaires were sent, addressed to the CEOs. Seven hundred thirty-six (736) questionnaires were received, 52 percent of which were answered by the CEO and 33 percent by an employee directly answering to the CEO, resulting in an 18 percent response rate. To check for sample selection bias by size and sector, we compare the response rates for the eighteen strata. There does not appear to be a serious sample selection bias by either size or sector. Whether selection is biased by the respondent's attitude towards HRM cannot be determined. (3)
Description and Construction of Variables
The professional HRM practices scale was developed from a subset of questionnaire items on recruitment practices, selection methods and procedures, compensation, training and development, and appraisal. Each of these items was measured on a three-point scale (1 = no; 2 = for some vacancies/jobs; 3 = for all vacancies/jobs). A list of all items can be found in the Appendix. The selected items all represent practices that are considered to be in accordance with professional standards and/or published as "best practices" for approaching that particular area based on judgments by a multinational group of experts from Australia and the Netherlands. The professional HRM practices scale was created in three steps. First, a separate factor analysis was carried out using Principal Components Analysis and a Varimax rotated solution to identify relevant items for each of the five categories of HRM practices (recruitment, selection, compensation, training and development, and appraisal). Second, the items selected for inclusion for each category were averaged to construct a separate subscale. To determine the reliability of these subscales, Cronbach's alpha reliability coefficients were calculated for the selected items. Finally, the professional HRM practices scale was calculated as an unweighted average of the underlying subscales.
Indicators of organizational complexity (and resource availability) include firm size, HRM specialization, formal business plan, and export strategy. Firm size was measured as the log (number of employees), including employees with temporary contracts, with no correction for part-time work. To measure the HRM specialization variable, respondents were asked two questions: whether or not an HRM department was present; and whether or not an HRM manager was present in the firm. These questions were then used to construct a dichotomous variable "HRM specialization" where 0 = neither HRM department nor HRM manager is present; 1 = either an HRM department or HRM manager is present (or both). To measure the formal business plan variable, respondents were asked whether or not a formal business plan or strategic plan exists (0 = no; 1 = yes). To measure export strategy, respondents were asked whether or not the firm exports (0 = no; 1 = yes).
The variable family firm was constructed as follows: a company received a score of 1 when it answered in the affirmative to both of the following questions: (1) members of one family own this business, and (2) members of one family manage this business. No distinction was made between firms with single owner-managers and those in which two or more family members own and/or manage the firm. It received a score of 0 otherwise. Finally, certain control variables were measured, including firm age; sector (service sector, trade sector, and manufacturing sector); franchising (0 = no; 1 = yes); and the percentage of employees belonging to a union (referred to as unionization in the following discussion).
For H1, two protocols were used to test for mediating effects of the resource availability indicators; one proposed by James and Brett (1984) and the second by Baron and Kenny (1986). According to Baron and Kenny (1986), one can test for the mediating effect of variable m (resource availability), by first examining the relationship between proposed antecedent x (family firm) and consequence y (professional HRM practices), and then investigating the extent to which this relationship diminishes (or even vanishes) if mediating variable, m is included in the model. The first step is to use bivariate tests to check for the significance of the relationships between x and y, x and m, and m and y. Second (given significant bivariate relationships), the models y = f(x) and y = f(m, x) must be estimated. To support the inference that m completely mediates the effect of x on y, the effect of x on y should be significant in the model y = f(x) but not in the model y = f(m, x).
Based on the same starting premise of significant bivariate relationships between x and y, x and m, and m and y, James and Brett (1984) compared the models y = f(m) and y = f(m, x). If the added effect of x (tested by the significance of the R-squared change when x is added to the first model) is not significant, m can be seen as completely mediating the relationship between x and y. Conversely, a significant result provides support for a direct effect.
In this study, we combine the two protocols by estimating three separate models: y = f(x), y = f(m) and y = f(x, m). We assume the presence of a mediating effect when the following requirements are met: (1) significant effect of m on y in the model y = f(m); (2) a significant effect of x on y in the model y = f(x); and (3) a nonsignificant effect of x on y in the model y = f(m, x). Likewise, we assume the presence of a direct effect in the case of a significant effect of x on y in the model y = f(x) in combination with a significant added effect of x on y in the model y = f(m, x).
A limitation of the just-described protocols is that the relationship between the family firm and the indicators of resource availability are established by looking at bivariate correlations. Because we hypothesize that the family firm variable is related to indicators of resource availability other than firm size, we elaborate on the protocol as follows: we control for firm size bias by estimating logistic regressions where the other indicators of resource availability (HRM specialization, formal business planning, and export strategy) are related to the family firm variable as well as firm size.
Scale Formation for Professional HRM Practices
The average scores, percentiles, and reliabilities of the professional HRM practice subscales are presented in Table 1.
For four of the five subscales (all but appraisal), Cronbach's alpha exceeds 0.60 (Table 1). According to criteria proposed by Nunnally (1978), (4) the reliabilities for these subscales are acceptable for an explorative study. The reliabilities of these subscales are comparable with those reported by Huselid (1995) and Delery and Doty (1996). With a Cronbach's alpha of 0.43, the reliability of the appraisal subscale is unsatisfactory. Given the importance of this subscale, we nevertheless decided to include it in our study.
Because of missing data, none of the subscales can be calculated for all firms. This introduces the risk of a selection bias. To determine whether such a bias may occur, we examined for each subscale and whether the respondents to that subscale differ significantly in their average scores on a number of variables compared to the nonrespondents. (5) No significant differences in firm size were found between respondents and nonrespondents. What does matter is the position of the respondent within the organization. For CEOs, the response rate is significantly lower than for other respondents. (6) This holds for all subscales, with the exception of the recruitment subscale. A possible explanation for this finding is that the CEO takes less time to fill in the complete questionnaire.
The professional HRM practices scale is calculated as an unweighted average of the underlying subscales. The resulting overall HRM scale is defined for 519 enterprises (Cronbach's alpha equals 0.78).
Descriptive and Bivariate Statistics
Table 3 reports the means, standard deviations, and correlation coefficients between the major variables in this study. The relationships between each of the organization characteristics variables and professional HRM practices are all expected to be positive. Reviewing the bivariate correlation statistics presented in Table 2 provides support for the relationship between professional HRM practices and all four organization characteristics variables, including firm size (r = 0.41; p < .01), HRM specialization (r = 0.43; p < .01), formal business plan (r = 0.35; p < .01), and export strategy (r = 0.19; p < .01).
The bivariate correlations between family firm and three of the four organization characteristics variables are fairly strong, statistically speaking, and negative (with firm size, r = -0.27, p < .01; with HRM specialization, r = -0.29, p < .01; with formal business plan, r = -0.24, p < .01), The correlation between the family firm variable and export strategy is somewhat weaker than for the other three relationships though still statistically significant at the 0.05 level (r = -0.08; p < .05).
To control for firm size bias, we estimate three logistic regressions, relating three indicators of resource availability (HRM specialization, formal business plan, and export strategy) to family firm as well as firm size. In addition, we have included sector dummies as control variables (Table 3). The results provide partial support for H1: even after controlling for firm size, organizational complexity (as indicated by HRM specialization and formal business plan) tends to be lower for family firms as compared to nonfamily firms. Only for export strategy do we find that the relationship with the family firm variable is no longer significant, once we control for firm size.
Finally, we note that the bivariate relationship between family firm and professional HRM practices is significant and negative (r = -0.40; p < .01).
The logistic regressions show the significance and sign of the relationships between the family firm variable (x), resource availability indicators (m), and professional HRM practices (y). We can now test for the mediating effect of resource availability in the relationship between family firm and professional HRM practices. The estimation results of the regression models for y = f(m, c), y = f(x, c) and y = f(m, x, c) are presented as models 1, 2, and 3 in Table 4, where c represents additional control variables in the equation. In addition, the last column in Table 4 reports the change in [R.sup.2] for two separate analyses (either when a block is entered first--without the control variables; or last in the all-variable regression model).
In a regression that only includes family firm as the independent variable, the estimated parameter equals -0.23 (p < .01), and this family firm effect explains approximately 15 percent of the variation in professional HRM practices.
Which part of this total family firm effect is mediated by firm size and indicators of organizational complexity? This can be determined by looking at the third model reported in Table 4. In this full model, the estimated family firm parameter is -0.14 (p < .01). The contribution to the [R.sup.2] when family firm is entered last in the equation, though reduced in magnitude, is still statistically significant ([DELTA][R.sup.2] = 0.05; p < .01). We therefore accept H2, and conclude that a direct effect of the family firm variable on professional HRM practices does exist.
In addition, we also accept H1, concerning the presence of an indirect effect of the family firm variable on professional HRM practices with indicators of firm size and organizational complexity serving as intervening variables. This conclusion follows from the findings that, first of all, family firm is related to each of these indicators; second, that these indicators have a significant effect on professional HRM practices (in the first as well as in the third model in Table 4); and finally, that the estimated family firm parameter decreases (from -0.23 to -0.14) when the effects of the organizational complexity indicators are added into the linear regression model.
Our results support both hypotheses set forth; namely that firms with family ownership and/or management are less likely to use professional HRM practices, and that this may be due both to direct and indirect effects of the independent variable. The direct effect can be explained by agency theory and the indirect effect by the resource-based view. In the discussion, we examine these premises more closely and discuss alternative theories that could also explain our results.
One explanation for the direct effect is offered by agency theory, according to which the family firm effect is due in part to a decreased (perceived or actual) need for monitoring of the management by the owner and of the employees by the management. However, organization control theory and institutional theory might also support a direct effect. Organizational control theory points out that clan and social control systems are more effective than the bureaucratic and administrative systems when strategy, decision-making, and power in the organization are managed by few people who share common values and coordinate themselves by informal relationships (Gnan and Songini 2003). It could be that in family firms, the social interactions among family members allow the use of informal and cultural mechanisms that substitute or complement the formal administrative systems.
Institutional theory may also help to explain the direct effect. Whereas agency theory focuses on the relationships between two specific stakeholders of an organization, institutional theory typically examines additional stakeholders. Institutional theorists view organizations as entities that gain legitimacy and stakeholder acceptance by conforming to their stakeholders' expectations. Examples of stakeholders are governmental institutions, professional organizations, and certifying bodies (Paauwe 1998).
Williamson (2000) uses institutional theory to develop a strategic model of small business recruitment. In particular, he introduces the notion of employer legitimacy, defined as "a generalized perception or assumption held by job applicants that an organization is a desirable, proper or appropriate employer given the system of norms, values, beliefs and definitions that exist within an industry" (Williamson 2000, p. 28). Williamson (2000) posits that to the extent that an organization's recruitment procedures and other human resource policies are viewed as proper and appropriate by potential job applicants, the organization will be seen as a legitimate employer. That is, small firms copy the HRM practices of larger firms to gain employer legitimacy. But within this context, one might argue that the family firm has less need of legitimacy to the extent that it attracts family and friends to work for the firm. In addition, employer legitimacy in the family firm may derive less from the professional manner in which a firm handles its HRM policies than from ways in which family ties are managed.
In addition, family ownership is associated with a desire to remain independent and keep full control over the organization (Bacon et al. 1996; Blais and Toulouse 1990). Case studies suggest that employers often associate professional HRM practices with a loss of control over (and flexibility of) the employee relations (Koch and De Kok 1999). This would provide an additional explanation for a direct negative effect of family ownership and management on professional HRM practices.
More than half of the family firm effect is explained by firm size and indicators of organization complexity. This is in line with the resource-based view of organizations. In particular, it is posited that these variables are likely to reflect greater resource availability and/or organization capabilities within the firm, making it easier for the firm to develop professional expertise in HRM practices as well.
Alternatively, these findings can be explained by the company growth theory, which suggests that as a company gets larger, the management task becomes more complex and requires a more professional approach. Rather than not being able to use professional HRM practices (due to a lack of resources), this argument suggests that professional HRM practices are less relevant to family firms.
The primary purpose of this study is to examine and explain differences in professional HRM practices between family and nonfamily-owned and managed firms. To this end, we have developed a model concerning direct and indirect effects of family firm characteristics on the use of professional HRM practices.
We find that, based on our sample and model, family firms are less likely to use professional HRM practices than their counterparts. This family firm effect occurs not only indirectly (since family businesses tend to be smaller, and/or less complex than nonfamily businesses, where complexity stimulates the application of professional HRM practices), but also directly. The direct effects are consistent with predictions consistent with agency theory, which predicts less monitoring in the family firm. On the other hand, the results cannot rule out other interpretations offered by organization control theory and/or institutional theory. Furthermore, the indirect family firm effect is consistent with predictions based on the resource-based view although once again, alternative interpretations of the findings cannot be ruled out.
Lacking performance data, we have not examined whether it is actually better or worse for family firms to rely upon less professional HRM practices. Our results suggest that future research into the relations between professional HRM practices and performance for small firms should include family ownership and/or management as a contingency variable.
Aldrich, H., and N. Langton, 1997, "Human Resource Management Practices and Organizational Life Cycles," in Frontiers of Entrepreneurship Research. Ed. P. D. Reynolds, W. D. Bygrave, N. M. Carter, R Davidsson, W. B. Gartner, C. M. Mason, and P. P. McDougall. Wellesley, MA: Babson College Center for Entrepreneurship, 349-357.
Audretsch, D. B., and A. R. Thurik (2001). "What Is New about the New Economy: Sources of Growth in the Managed and Entrepreneurial Economies," Industrial and Corporate Change 10, 267-315.
Bacon, N., P. Ackers, J. Storey, and D. Coates (1996). "It's a Small World: Managing Human Resources in Small Business," The International Journal of Human Resource Management 7(1), 82-100.
Baron, R. M., and D. A. Kenny (1986). "The Moderator-Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations," Journal of Personality and Social Psychology 51(6), 1173-1182.
Blais, R. A., and J.-M. Toulouse (1990). "National, Regional or World Patterns of Entrepreneurial Motivation?" Journal of Small Business and Entrepreneurship 7(2), 3-20.
Boxall, P. (1996). "The Strategic HRM Debate and the Resource-Based View of the Firm," HRM Journal 6(3), 59-75.
Cappelli, P., and A. Crocker-Hefter (1996). "Distinctive Human Resources Are Firms' Core Competencies," Organizational Dynamics 24(3), 7-22.
Chandler, G. N., and G. M. McEvoy (2000). "Human Resource Management, TQM, and Firm Performance in Small and Medium-Size Enterprises," Entrepreneurship Theory and Practice 25(1), 43-58.
Cromie, S., B. Stephenson, and D. Monthieth (1995). "The Management of Family Firms: An Empirical Investigation," International Small Business Journal 13(4), 11-34.
Cyr, L. A., D. E. Johnson, and T. M. Welbourne (2000). "Human Resources in Initial Public Offering Firms: Do Venture Capitalists Make a Difference?" Entrepreneurship Theory and Practice 25(1), 77-91.
Daily, C. M., and M. J. Dollinger (1993). "Alternative Methods for Identifying Family-Versus Nonfamily-Managed Businesses," Journal of Small Business Management 31(2), 79-90.
Damanpour, F. (1996). "Organizational Complexity and Innovation: Developing and Testing Multiple Contingency Models," Management Science 42(5), 693-716.
De Kok, J. M. P. (2003). "Human Resource Management within Small and Medium-Sized Enterprises," Tinbergen Institute Research Series 313. Amsterdam: Thela Thesis.
De Kok, J. M. P., and L. M. Uhlaner (2001). "Organizational Context and Human Resource Management in the Small Firm," Small Business Economics 17(4), 273-291.
De Kok, J. M. P., L. M. Uhlaner, and A. R. Thurik (2003). "Human Resource Management with Small and Medium-Sized Enterprises: Facts and Explanations," ERIM Report Series ERS 2003-015, Rotterdam: Erasmus University.
Delery, J. E., and D. H. Doty (1996). "Modes of Theorizing in Strategic Human Resource Management: Tests of Universalistic, Contingency and Configurational Performance Predictors," Academy of Management Journal 4(39), 802-835.
Dyer, W. G. (2003). "The Family: The Missing Variable In Organizational Research," Entrepreneurship Theory and Practice 27(4), 401-416.
Fiegener, M. K., B. M. Brown, R. A. Prince, and K. M. File (1996). "Passing on Strategic Vision," Journal of Small Business Management 34(3), 15-26.
Flanagan, D. J., and S. P. Deshpande (1996). "Top Management's Perceptions of Changes in HRM Practices after Union Elections in Small Firms," Journal of Small Business Management 34(4), 23-34.
Gnan, L., and L. Songini (2003). "The Professionalization of Family Firms: The Role of Agency Cost Control Mechanisms," FBN Proceedings 2003, 141-172.
Golhar, D. Y., and S. P. Deshpande (1997). "HRM Practices of Large and Small Canadian Manufacturing Firms," Journal of Small Business Management 35(3), 93-97.
Gomez-Mejia, L. R., M. Nunez-Nickel, and I. Gutierrez (2001). "The Role of Family Ties in Agency Contracts," Academy of Management Journal 44(1), 81-95.
Hendry, C., A. Jones, M. Arthur, and A. Pettigrew (1991). Human Resource Development in Small to Medium Sized Enterprises. London: Employment Department.
Heneman, H. G., and R. A. Berkley (1999). "Applicant Attraction Practices and Outcomes among Small Businesses," Journal of Small Business Management 37(1), 53-74.
Heneman, R. L., J. W. Tansky, and S. M. Camp (2000). "Human Resource Management Practices in Small and Medium-Sized Enterprises: Unanswered Questions and Future Research Perspectives," Entrepreneurship Theory and Practice 25(1), 11-26.
Huselid, M. A. (1995). "The Impact of Human Resource Management Practices on Turnover, Productivity, and Corporate Financial Performance," Academy of Management Journal 38(3), 635-672.
Jackson, S. E., R. S. Schuler, and J. C. Rivero (1989). "Organizational Characteristics as Predictors of Personnel Practices," Personnel Psychology 42, 727-786.
James, L. R., and J. M. Brett (1984). "Mediators, Moderators, and Tests for Mediation," Journal of Applied Psychology 69(2), 307-321.
Jorissen, A., E. Laveren, R. Martens, and A. Reheul (2002). "Differences between Family and Nonfamily firms: The Impact of Different Research Samples with Increasing Elimination of Demographic Sample Differences," Conference Proceedings, RENT XVI, 16th workshop, Nov. 21-22, 2002, Barcelona. Spain: Univeristat Autonoma de Barcelona.
Julien, P., and C. Ramangalahy (2003). "Competitive Strategy and Performance of Exporting SMEs: An Empirical Investigation of the Impact of Their Export Information Search and Competencies," Entrepreneurship Theory and Practice 27(3), 227-245.
Klaas, B. S., J. McClendon, and T. W. Gainey (2000). "Managing HR in the Small and Medium Enterprise: The Impact of Professional Employer Organizations," Entrepreneurship Theory and Practice 25(1), 107-124.
Koch, C.L.Y., and J. M. P. De Kok (1999). "A Human Resource-Based Theory of the Small Firm," Research Report 9906/E, Zoetermeer. The Netherlands: EIM Small Business Research and Consultancy.
Lawler, E. E., and G. D. Jenkins (1992). "Strategic Reward Systems," in Handbook of Industrial and Organizational Psychology, Vol. 3, 2nd ed. Ed. M. D. Dunnette and L. M. Hough. Palo Alto, CA: Consulting Psychologists Press, 1009-1055.
Matlay, H. (1999). "Employee Relations in Small Firms: A Micro-Business Perspective," Employee Relations 21(3), 285-295.
Meyer, A. D., A. S. Tsui, and C. R. Hinings (1993). "Introduction: Configurational Approaches to Organizational Analysis," Academy of Management Journal 36, 1175-1195.
Michalisin, M. D., R. D. Smith, and D. M. Kline (1997). "In Search of Strategic Assets," International Journal of Organizational Analysis 5, 360-387.
Nunnally, J. C. (1978). Psychometric Theory, 2nd ed. New York: McGraw-Hill.
Paauwe, J. (1998). "HRM and Performance: The Linkage between Resources and Institutional Context," RIBES Paper 9854. Rotterdam, the Netherlands: Erasmus University Rotterdam.
Pfeffer, J. (1994). Competitive Advantage Through People. Boston: Harvard Business School Press.
Pollak, R. (1985). "A Transaction Cost Approach to Families and Households," Journal of Economic Literature 23(2), 581-608.
Priem, R. L., and J. E. Butler (2001). "Is the Resource-Based 'View' A Useful Perspective for Strategic Management Research?" Academy of Management Review 26(1), 22-40.
Randoy, T., and S. Goel (2003). "Ownership Structure, Founder Leadership, and Performance in Norwegian SMEs: Implications for Financing Entrepreneurial Opportunities," Journal of Business Venturing 18, 619-637.
Reid, R. S., and J. S. Adams (2001). "Human Resource Management--A Survey of Practices within Family and Non-Family Firms," Journal of European Industrial Training 25(6), 310-320.
Schermerhorn, J. R. (2001). Management Update 2001, 6th ed. New York: John Wiley and Sons, Inc.
Schulze, W. G., M. H. Lubatkin, and R. N. Dino (2003). "Toward a Theory of Agency and Altruism in Family Firms," Journal of Business Venturing 18, 473-490.
Sirmon, D. G., and M. A. Hitt (2003). "Managing Resources: Linking Unique Resources, Management and Wealth Creation in Family Firms," Entrepreneurship Theory and Practice 27(4), 339-358.
Steier, L. (2003). "Variants of Agency Contracts in Family-Financed Ventures as a Continuum of Familial Altruistic and Market Rationalities," Journal of Business Venturing 18, 597-618.
Wagar, T. H. (1998). "Determinants of Human Resource Management Practices in Small Firms: Some Evidence from Atlantic Canada," Journal of Small Business Management 36, 13-23.
Wiesner, R., and J. McDonald (2001). "Bleak House or Bright Prospect? Human Resource Management in Australian SMEs," Asia Pacific Journal of Human Resources 39(2), 28-50.
Williamson, I. O. (2000). "Employer Legitimacy and Recruitment Success in Small Businesses," Entrepreneurship Theory and Practice 25(1), 27-42.
Subscales of Formal HRM Practices
This appendix provides additional information on the HRM subscales used in this study. The HRM subscales are defined by a selection of available items in the questionnaire. Each of these items is measured on a three-point scale (no, for some vacancies/jobs, for all vacancies/jobs). The following table presents the selected items that are used to define the subscales. An elaborate discussion of the subscales, including a discussion of the correspondence with the results of factor analysis on all items, can be found in De Kok, Uhlaner, and Thurik (2003).
Items Included in Subscales Subscale Item Recruitment Recruitment and selection office Temporary employment agencies Magazines Internet Referrals by employees References from other sources Open house Selection Use of written job descriptions Job analysis Psychological tests Interview panels Compensation Performance pay (Partly) based on job evaluation Competitive wages Wages based on acquired skills Group incentive programs Individual incentive programs Profit sharing Annual bonus Additional financial benefits, other than pensions (for example, insurance and savings arrangements) Training and Training provided to employees development Formal training budget available Recent introduction of formal training programs Recent intensification of existing training programs Formal in-house training by internal staff Formal in-house training by external staff External training Management and development training Appraisal Rating scales Management by objectives Appraisal conducted by line manager
*The present report has been written in the framework of the research program SCALES which is carried out by EIM and financed by the Dutch Ministry of Economic Affairs. Lorraine Uhlaner also acknowledges financial support of Arenthals Grant Thornton Netherlands, Fortis Bank, and Mees/Pierson Bank, a subsidiary of Fortis Bank specialized in private wealth management. An earlier version of this paper has been presented at the 2001 RENT conference on research in entrepreneurship (Turku, Finland).
Jan M. P. de Kok is senior researcher at EIM.
Lorraine M. Uhlaner is director, European Family Business Institute, Erasmus School of Economics and Business Economics.
A. Roy Thurik is professor of economics and entrepreneurship, Erasmus School of Economics and Business Economics; director, Centre for Advanced Small Business Economics (CASBEC); and research professor of entrepreneurship, growth and public policy, Max Planck Institute of Economics in Jena (Germany).
Address correspondence to: Jan de Kok, EIM Business & Policy Research, P. O. Box 7001, 2701 AA Zoetermeer, The Netherlands. Tel: +31(0)79 343-0601. E-mail: email@example.com.
(1) The terms "family-owned and managed firms" and "family firms" are used as synonyms in this paper.
(2) Even when controlling for size and other factors.
(3) The details of sampling can be found in De Kok, Uhlaner, and Thurik (2003).
(4) For early stages of basic research, Nunnally (1978) suggested that a Cronbach's alpha between 0.5 and 0.6 would be sufficient.
(5) These control variables are size, sector, current working position of the respondent, location of the firm, current tenure of the respondent, whether the respondent is (part) owner, whether the company is owned by a family, whether the enterprise is member of a franchise organization, if a business plan is available, and the respondent's gender.
(6) The response rate is also lower if the respondent has a long tenure with the firm, or is (part) owner of the firm. Because ownership, tenure, and being CEO are strongly related with each other, these differences in response rate are interpreted as confirmations of the CEO-effect.
Table 1 Scores and Reliability Statistics on Subscales of Professional HRM Practices (a) Recruitment Selection Compensation Score Mean 1.6 2.0 1.8 10 percent 1.3 1.5 1.3 percentile 90 percent 2.0 2.8 2.2 percentile Reliability Cronbach's 0.62 0.69 0.64 alpha Valid 533 619 621 Observations Training and Development Appraisal Score Mean 1.7 2.0 10 percent 1.0 1.3 percentile 90 percent 2.3 2.7 percentile Reliability Cronbach's 0.81 0.43 alpha Valid 669 598 Observations (a) All subscales are defined on the interval [1,3]. Table 2 Pearson Correlation between All Variables for the Total Sample 1 2 3 4 5 1. Family Firm (a) 1 2. Firm Size (log -0.27** 1 employees) 3. Formal Business -0.24** 0.25** 1 Plan (0 = no; 1 = yes) (a) 4. Franchise (0 = no; 0.05 -0.01 -0.00 1 1 = yes) (a) 5. Unionization (%) -0.03 0.13** 0.07 -0.10** 1 6. HRM Specialization -0.29** 0.41** 0.22** -0.06 0.05 7. Professional HRM -0.40** 0.41** 0.35** 0.03 0.03 Practices 8. Firm Age (Log 0.06 0.11** -0.03 -0.20** 0.17** [firm age]) 9. Export -0.08* 0.10** 0.06 -0.11** 0.09* Strategy (a) 10. Trade Sector (a) -0.00 0.00 -0.03 0.17** -0.16** 11. Manufacturing 0.01 0.08* 0.04 -0.12** 0.29** Sector (a) 12. Service Sector (a) 0.00 -0.11** -0.02 0.03 -0.19** Mean 0.45 4.16 0.66 0.07 1.52 S.D. 0.50 0.79 0.47 0.25 1.33 0.50 N 695 700 685 671 670 6 7 8 9 10 1. Family Firm (a) 2. Firm Size (log employees) 3. Formal Business Plan (0 = no; 1 = yes) (a) 4. Franchise (0 = no; 1 = yes) (a) 5. Unionization (%) 6. HRM Specialization 1 7. Professional HRM 0.43** 1 Practices 8. Firm Age (Log -0.01 -0.03 1 [firm age]) 9. Export 0.12** 0.19** 0.06 1 Strategy (a) 10. Trade Sector (a) -0.03 -0.04 0.05 0.09* 1 11. Manufacturing -0.03 0.06 0.07 0.25** -0.28** Sector (a) 12. Service Sector (a) 0.00 -0.06 -13** -0.32** -0.33** Mean 0.54 1.82 3.37 0.27 0.13 S.D. 0.50 0.30 1.00 0.44 0.33 0.48 N 696 519 682 693 695 11 12 1. Family Firm (a) 2. Firm Size (log employees) 3. Formal Business Plan (0 = no; 1 = yes) (a) 4. Franchise (0 = no; 1 = yes) (a) 5. Unionization (%) 6. HRM Specialization 7. Professional HRM Practices 8. Firm Age (Log [firm age]) 9. Export Strategy (a) 10. Trade Sector (a) 11. Manufacturing 1 Sector (a) 12. Service Sector (a) -0.62** 1 Mean 0.35 0.42 S.D. 0.50 0.49 N 695 695 (a) Dummy variable. The relation between two nominal variables is measured by the Phi coefficient. For dummy variables, the Phi coefficient is identical to Pearson's correlation. Fisher's exact test was used to test for dependency between two dummy variables. *p < .05-level (2-tailed). **p < .01-level (2-tailed). Table 3 Results of Logistic Regression to Examine the Relationship between Family Ownership and Three Indicators of Organizational Complexity Organizational Characteristic (d) HRM Formal Export Variable Specialization Business Plan Strategy Family Firm -0.78** -0.75** -0.32 Firm Size 1.26** 0.65** 0.22 Firm Age -0.11 -0.11 0.06 Trade Sector -0.02 -0.21 -0.22 Service Sector 0.34 -0.09 -1.77** Other Sector 0.62* -0.12 -0.49 Goodness of Fit Measures % Predicted Correctly (a) 70 (54) 71 (67) 72 (73) [R.sup.2] (Nagelkerke) 0.29 0.14 0.17 [chi square] Test for Model 7.79 (0.099) 2.2 (0.70) 71 (0.00) Parsimony (b) [chi square] Test for Model 161 (0.00) 68 (0.00) 81 (0.00) fit (c) Valid Observations 669 660 668 (a) The reference value (the share of firms with HRM specialization/ formal business plan/export strategy) is reported within parentheses. (b) Test for the joint hypothesis that the parameters for age, trade, service and other sectors are equal to zero. Probability value within parentheses. (c) Test for the hypothesis that all included parameters (except constant) are equal to zero. Probability value within parentheses. (d) The significance of the parameters is based upon both Wald statistics and Likelihood Ratio test statistics. Both test statistics lead to the same conclusions. A constant term has been estimated, but is not included in the table. *Significant at 5% level. **Significant at 1% level. Table 4 Results of Ordinary Least Squares Regression Analyses on Professional HRM Practices (b) Model 1 Controls and Organization Model 2 Controls Characteristics and Family Firm Explanatory Variables B-Value t-Value B-Value t-Value Organization Characteristics Firm Size (Log) 0.09 5.16** Formal Business Plan 0.14 5.51** Export Strategy 0.09 3.26** HRM Specialization 0.16 6.11** Family Firm -0.23 -9.05** Controls Firm Age (log) -0.01 -0.58 0.00 0.19 Unionization (%) 0.01 -1.51 0.00 -0.30 Manufacturing Sector 0.05 1.67 0.04 1.03 Service Sector 0.01 0.26 -0.02 -0.70 Franchise 0.09 1.65 -0.23 -9.05 Constant 1.25 15.86** 1.91 35.72** R-Square 0.31 0.16 Adjusted R-Square 0.30 0.15 F-Statistic 22.86** 14.56** Model 3 All Variables Explanatory Variables B-Value t-Value [DELTA][R.sup.2] (a) Organization Characteristics 0.30**/0.20** Firm Size (Log) 0.07 4.40** Formal Business Plan 0.13 5.21** Export Strategy 0.08 3.06** HRM Specialization 0.13 5.28** Family Firm -0.14 -6.01** 0.15**/0.05** Controls 0.01/0.01 Firm Age (log) 0.00 0.14 Unionization (%) -0.01 -1.62 Manufacturing Sector 0.05 1.62 Service Sector 0.00 0.15 Franchise 0.09 1.82 Constant 1.38 17.46** R-Square 0.36 Adjusted R-Square 0.35 F-Statistic 25.78** (a) Change in [R.sup.2] when adding this variable first/last to the model (including all variables, that is, organization characteristics, family firm, controls). (b) B-values refer to the unstandardized coefficients of the explanatory variables. *p < .01-level. **p < .001-level. #p < .05.…