Acquiring versus growing star performers: The effects on
organizational performance
Master Thesis
Student: Hans Schut / Student No 10383387 /
MSc. Business Administration, Strategy track
University of Amsterdam, Faculty of Economics and Business Supervisor: Dr. N.E. (Nathan) Betancourt
University of Amsterdam, Amsterdam Business School Date: June 22nd 2018 final version
Statement of originality
This document is written by student Hans Schut who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its
references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Table of Contents
ABSTRACT ... 4
INTRODUCTION ... 5
LITERATURE REVIEW AND HYPOTHESES ... 7
STAR PERFORMERS ... 8
INTERNALLY DEVELOPED STARS... 10
THE MODERATING EFFECT OF HOMOGENEOUS EXPERTISE ... 12
METHODOLOGY ... 14
SAMPLE AND DATA COLLECTION ... 16
MEASURES……….. ... 17 Dependent Variables ... 17 Failure ... 18 Growth ... 19 Independent variables ... 21 Moderator variable ... 23 Control variables ... 24
RESULTS OF THE MULTIVARIATE ANALYSIS ... 25
DESCRIPTIVE STATISTICS AND CORRELATION ANALYSIS ... 26
REGRESSION ANALYSIS... 29
DISCUSSION ... 36
MAJOR FINDINGS ... 37
CONTRIBUTIONS ... 40
LIMITATIONS AND FUTURE RESEARCH ... 41
CONCLUSION ... 43
Abstract
The idea of acquiring star performers to gain a competitive advantage has resulted in a war for talent in knowledge intensive industries. However, scholars seem to disagree on the potential effects of stars on organizational performance. By using a dataset that has followed the movement of international law firm partners in Hong Kong, over an eleven-year period, this thesis has tested whether star performers that are internally developed or externally hired have different effects on firm performance. Furthermore, it has been tested if adding a star with a similar expertise to an organization that already consists of stars, can moderate the
relationship between organizational performance and the adding of stars. Thereby, this thesis contributes to the literature regarding the effects of star performers on organizational
performance. To test the formulated hypotheses, the data has been analysed using a Probit-regression to cope with a dependent binary variable and Linear-Probit-regressions to cope with continuous dependent variables. The results found in this research show that for this specific dataset internally developed stars are more likely to improve firm performance than externally hired stars. Furthermore, the hypothesized moderating effect of a star’s expertise match is not supported by the results. Moreover, this research notes that more research is needed on this topic for different industries to be able to generalize the results across industries and countries.
Introduction
A firm’s success and future survival depends on the quality of its human capital (Kehoe & Tzabbar, 2014). As new knowledge intensive industries have changed the focus from machines to people; the human capital of organizations has now become the source of competitive advantage. The concept of hiring the best talent to boost organizational
performance is not new and was originally presented by McKinsey & Company in their 1998 report “Better talent is worth fighting for” (Chambers et al., 1998). This research will focus on the effects of top performing talent, on organizational performance.
The top performers in organizations that are discussed in this thesis are referred to as “stars”. Star employees, are defined as individuals who demonstrate exceptionally high productivity (Groysberg, lee & Nanda, 2008), and thereby are very valuable to organizations. Their superior production compared to their peers is believed to have positive effects on firm performance (Zucker & Darby, 1997). The positive effects accompanied by the acquirement of a star can be drastic, so much even that the addition of a star can lead to a firm’s success, and the departure of a star can signal its decline or even failure (Aguinis & O’Boyle JR., 2014). When the potential of star performers was recognized, an industry wide search for talent started. Chambers et al. (1998) predicted how the war for talent would intensify. And indeed, in recent years both Law and Tech Firms in the United States have indeed drastically increased their offers to talented associates and tech experts. Focusing on this idea of
acquiring stars to boost company performance, this thesis will focus on the effects of acquiring stars on organizational performance. One would expect an abundant supply of literature regarding the effects of the acquirement of stars on organizational performance. However, there is still a lacking understanding of star performers (Aguinis & O’Boyle Jr., 2014), and scholars seem to be in conflict regarding the effects of the acquirement of stars on organizational performance.
When a star is added to a group without stars this is found to have a positive effect on the opportunities of the group as the star enhances; the external visibility, the acquirement of resources, and talent and also the development of knowledge in the group (Groysberg, Polzer & Elfenbein, 2011; Olroyd & Morris, 2012). Grioriou & Rothaermel (2014), further note that firms that employ stars also enjoy several innovation-related benefits such as improved; knowledge sensing, renewal, knowledge capture, and a higher ability to adaptability to radical discontinuities. Agarwal, McHale, and Oettl (2013) lastly argue that the arrival of a star can be an instrument to attract more talent. These positive consequences of the acquirement and employment of stars can lead to organizations raising their performance.
However, there is also an increasing opposing argument noticeable in the literature. It draws attention to the considerable negative effects that accompany the acquiring of stars (Groysberg, Lee, Nanda 2008; Kehoe & Tzabbar, 2014; Groysberg, Nanda & Nohria 2004; O’Reilly & Pfeffer, 2000). Groysberg, Lee, & Nanda (2008) discuss the portability of the knowledge of stars and find that their performance drops drastically for up to 5 years after moving to another firm, because the portability of a star’s knowledge is often less than expected. Moreover, the hiring of a star could lead to a demotivated workforce (Groysberg, Nanda, & Nahria, 2004). There thus is a debate noticeable regarding the effectiveness of acquiring a star on organizational performance. This thesis proposes that whether a star is added or internally developed influences the effect of the star on the organizational
performance. Furthermore, the expertise of the star then moderates this relationship, and can help explain the possible difference in effect between the hiring and developing of a star on organizational performance.
The internal development of stars could give organizations all the benefits associated with star performers without the possible drawbacks of the acquirement of stars. However, the internal development of employees is a business that requires a commitment-based HR
system that builds employee involvement and many resources. Organizations will need to invest in training initiatives, career development, and mentor programs to build these skills (Lepak & Snell, 1999). The development of stars also requires significant resources before the organization can enjoy the benefits of an internally developed star. This stirs up the question: Does internally developing versus externally hiring stars have a different effect on
organizational performance?
To answer this question a dataset is used that observed partner movement, partner performance and firm performance of 180 law firms in Hong Kong, simultaneously over a period of 11 years beginning on January 1, 1998 and ending on December 31, 2008. With the help of this dataset, this thesis will first test if the movement of stars among law firms in Hong Kong will increase or hurt the acquiring firm’s performance. Secondly, it will examine if internal development of stars increases or hurts firm performance. Thirdly, the moderating effect of a star’s expertise on firm performance will be tested.
This research will add to the existing literature because there is still a lot of debate regarding the effectiveness of acquiring stars on firm performance. It could even be possible for a firm to choose specifically not to acquire a star, as the benefits do not outweigh the costs. This research will try to shed light on the effects of partner change on firm performance and hopes to resolve a small proportion of the argument regarding star performers.
Literature review and hypotheses
The following chapter reviews the main insights and arguments of the existing literature on star performers, organizational performance, and presents the hypotheses of this thesis. First, the definition, different concepts, and points of view regarding star performers will be discussed. Second, the internal development of stars will be discussed. Third, the potential
moderating effect of homogeneous expertise will be addressed. The hypotheses will be composed by combining the concepts of star performers and organizational performance. Star performers
Multiple definitions of star performers have been formulated in the past decades. The most used typology of a star is: “Star performers are individuals who demonstrate exceptionally high productivity and enjoy broad external visibility” (Groysberg, lee & Nanda, 2008; Olroyd & Morris, 2012). Stars that are well known, not only within but also outside the organization, because of their exceptional productivity, enjoy what is called broad external visibility. This can help the star and the organization with better finding and exploiting new opportunities. Because of a star’s high productivity, they are regarded as very valuable to organizations (Zucker & Darby, 1997).
Hiring a star can have several positive effects on the opportunities for the organization. The first is that the star enhances the external visibility of the organization which translates to being recognized by other talent, investors and clients who are more easily attracted due to their recognition of the exceptional performance of the star. This then can help the
organization with the acquirement of resources and development of knowledge in the group (Groysberg, Polzer & Elfenbein, 2011). The second advantage is that peers working with a star benefit from working with a talented colleague. This effect is the strongest in knowledge intensive industries (Groysberg, Polzer & Elfenbein, 2011). By working with stars, other members in the group broaden their skills and competencies, thereby increasing their
contributions to the group. Furthermore, when a firm hires a star in a certain department, the productivity of the department is found to raise with 26 percent (Agrawal, McHale, and Oettl, 2013). Secondly, they find a significant effect regarding the attraction of more talent to the organization. After the arrival of a star the measured quality of joining scientists has risen with 70 percent. External talents recognize the value of a star and are more willing to join
organizations where they can learn from these stars. Thirdly, they find that non-star scientists that work on related topics as the stars improve their output by almost 50 percent, in contrast to scientists that work on unrelated topics who present an increase of 10 percent (Agrawal, McHale, and Oettl, 2013). In addition to a star’s ability to improve performance of others and attract more talent, stars also seem to be better in capitalizing on their social networks.
The networks of relations are the social capital of the star, in which a higher number of relations or sources of information translates to more social capital. These networks of
relationships are part of the valuable resources a star possesses (Olroyd & Morris, 2012). Stars are visible within the company and on the labour market because of their exceptional productivity, leading to other peers and individuals trying to establish relationships with them. This in turn leads to exceptional high levels of social capital of the star (Olroyd & Morris, 2012). Grioriou & Rothaermel (2014), find that stars do not only identify new opportunities for organizations but also select the most promising ones. By using their network and relations intelligently, they can combine networks of knowledge that were not visible to others in the organization. This will then increase the performance of the star and the
individuals working with him/her (Grioriou & Rothaermel, 2014). To bring these advantages into a more abstract perspective, stars can be seen as a “VRIN” resource described by Barney (1991) in his article about gaining competitive advantages. The acquisition of a star is not just the filling of a position in the organization, but an investment in a resource that enables the organization to gain a competitive advantage. It is no surprise then that stars are in high demand by organizations searching for competitive advantages. This idea has gained much support and has been popularized over the past decades by management experts, resulting in a war for talent (Groysberg, 2010). Taking the idea of acquiring stars to boost company
performance the following hypothesis has been formulated:
Internally developed stars
As previously mentioned, there is also an increasing opposing argument in the literature (Groysberg, Lee, Nanda 2008; Kehoe & Tzabbar, 2014; Groysberg, Nanda & Nohria 2004; O’Reilly & Pfeffer, 2000). Groysberg, Lee, & Nanda (2008) discuss the portability of the knowledge of stars. In knowledge-work, star performance has traditionally been attributed to general human capital which is not firm specific (Groysberg, Lee, & Nanda, 2008). This assumes that the star is able to move freely between employers as they take their skills with them. However, Groysberg, Lee, & Nanda (2008) have presented in their findings that the performance of stars drops drastically for up to 5 years after the stars moves to a new firm. They assign this drop to multiple aspects.
First is the fact that the knowledge of a stars that leads to their performance is
embedded in firm specific human capital, therefore this is not portable. Stars depend on their colleagues and network within the firm; close collaboration with colleagues in project teams gives the star the edge (Groysberg, Lee, & Nanda, 2008). Second, the systems and practices embedded in organizations differ. A star that moves thus suffers a drop of performance due to difference in quality support. Third, a star has more information about his or her own skill than the hiring firm. Hiring firms are unsure about the workers abilities and will thus offer a wage based on expected performance. When this performance depends on firm specific human capital, the hiring firm is likely to overpay the star because he or she will not immediately be able use of the firm specific human capital sufficiently (Groysberg, Lee, & Nanda, 2008). Groysberg, Nanda, & Nahria (2004) share this view and further add that the filling of positions with externally hired stars can lead to the demotivation and demoralization of employees. They see resources and opportunities go to stars, which then incites the
assumption that they should look outside the organization if they were to aspire to the occupation of leadership positions. Furthermore, Groysberg, Nanda, & Nohria (2004) note
that stars find it harder to transition to a new organization because they are often handicapped by the attitudes of their new colleagues. Other managers and personnel who feel resentful of the high status and pay of the star will avoid them, cut off information and refuse to cooperate (Groysberg, Nanda, Nohria, 2004). At the same time, the star has to unlearn old habits and get accustomed to new procedures at the new firm. The combination of all these factors puts a strain on the performance of the star.
It is not only the management that can act resentful towards a new star. When a newly acquired star joins a group with other stars with similar expertise, a competition and fights for recognition and higher functions can start (Groysberg, Polzer & Elfenbein, 2011). Stars may deploy competitive tactics rather than seeking the best solution to win discussions and disputes because of their ego’s (Hambrick, 1994). Therefore, groups with a high number of stars can become dysfunctional and counterproductive (Groysberg, Polzer & Elfenbein, 2011). Kehoe & Tzabbar (2014) further argue that the resources and attention conferred to a star may limit opportunities of other, non-star, inventors to provide innovative leadership. The internal development of innovative leaders is important to firms as it makes the company less dependent on one, individual star for innovation. Firms should be cautious in conferring a large part of their resources to the hiring of stars because they tend to not stay long within one specific organization (Groysberg, Nanda, Nohria, 2004). This can leave the organization vulnerable.
A possible solution to these problems with externally hiring stars would be to develop the top talent in-house. The internal development of stars gives the organization all the benefits associated with star performers, but with less risk of the drawbacks of the acquirement of stars. Bidwell (2011) argues that firm have better information about their current employees, which allows them to better assess the qualities of the employee. This leads to lower risk of overpaying. When an employee is internally developed the work that the
employee carries out changes gradually, even when the title and the change of responsibilities of the job imply a bigger change (Bidwell, 2011). Furthermore, he finds that workers who are promoted into new jobs perform significantly better compared to their hired colleagues. This can be explained by the fact that externally hired employees must learn new skills before they can perform as well as the internally developed employees. Overall it seems that the internal development of employees is a more harmonious process compared to external hiring. Nonetheless, the externally hired employees earn 18 percent more on average (Bidwell, 2011).
Firms have to balance their decision based on the question if the benefits of internally developing a star outweigh the costs associated with the development of the star (Lepak & Snell, 1999). As mentioned earlier, the internal development of employees is a business that requires many resources and a well-organized and committed HR-system. Organizations will have to invest in extensive training initiatives, career development, and mentor programs to create these skills (Lepak & Snell, 1999).
Based on the insights into the possible negative consequences related to the acquirement of stars and the possibility of internally developing stars, the following hypotheses have been formulated:
H2: Internally developing a star has a positive effect on firm performance.
H3: Internally developing a star has a more positive effect on firm performance than externally hiring a star.
The moderating effect of Homogeneous Expertise
It has been discussed in literature how stars, working in groups with other stars in similar disciplines, can start to severely compete. Discussions about who has the right answer can lead to aggression, avoidance and the withholding of information when working on tasks.
(Groysberg, Polzer & Elfenbein, 2011; Overbeck et al., 2005; Tiedens & Fragale, 2003) However, this problem seems to be less severe in groups that have a heterogeneous pool of stars. Groysberg, Polzer and Elfenbein (2011) find that when stars can proclaim their expertise on certain content without threatening or challenging the expertise of others, the urge to compete stays lower. It seems that the competing behaviour of stars can restrain the collaborative behaviours and the outcome of the task (Groysberg, Polzer & Elfenbein, 2011), leading to lower performance. Performance of organizations is related to the social
interactions of its workforce. When organizational members have more social interactions such as coordination, trust and knowledge sharing, higher levels of performance are achieved (Huang & Li, 2009). These findings indicate that when stars work together, a higher
performance can be achieved compared to when they are working against each other. The latter is being supported by findings in the study of Sparrowe et al. (2001). They found support for the notion that negative relationships in groups, that lead to hindrance of work, lower group performance. These findings suggest that the negative effects observed in homogeneous groups of stars can lead to lower performance of the organization. Therefore, adding a star with a similar expertise to an organization that already consists of stars, can moderate the relationship between organizational performance and the effect of adding stars. Based on this insight the following hypotheses are formulated:
H4: Homogeneous expertise of a partner will decrease the positive effect of externally hired stars on firm performance.
H5: Homogeneous expertise of a partner will decrease the positive effect of internally developed stars on firm performance.
Methodology
This thesis aims to compare the effects of the acquirement of stars to that of the development of stars on organizational performance, by answering the following research question: Does growing stars have a different effect on organizational performance than acquiring them? Moreover, what is the effect of a star’s expertise match on this outcome? Based on the hypotheses set before, the following model is presented (See figure 1.)
Figure 1: Model of hypothesized relations. To answer the research question, that is the main focus of this thesis, multiple hypotheses have been formulated. The model presented in figure 1 shows the different hypothesized relations between the dependent variable Firm Performance, and the independent variables
External Hire of a star and Internal Development of a star. A star’s expertise match
moderates the relation of the independent variables.
Firstly, it is hypothesized that externally hiring a star has a positive effect on firm performance. Secondly, it is hypothesized that the internal development of a star has a positive effect on firm performance. Additionally, this effect is more positive than the effect of externally hiring a star. Lastly, the expertise of the star is expected to lower this positive effect, if the star’s performance is matched to the expertise of the firm. In this research the expertise match will be presented as Homogeneous Expertise.
To assess the presented model, a group of law firm partners is selected to test the hypothesis. Law firm partners are relevant subjects for this research for two reasons. First,
partners of law firms have proven to be the exceptionally productive individuals. They must outwork their peers for many years in order to make the step from associate to partner. Moreover, next to working under high stress and high competition, they also need to build social networks outside the firm that help them win over clients. Only associates that have managed to excel in both performance and the capitalization on social capital by bringing in new clients over a significant number of years have a chance to become partner with a firm. Therefore, the concept of a law firm partner is in line with the definition given by Groysberg, lee, & Nanda, (2008) and Zucker & Darby (1997), who emphasize the importance of their exceptional productivity compared to their peers. Secondly, law firm stars are a suitable research subject as there is a lot of mobility regarding the acquirement of talent, there is much development of talent, and because law firms are all similar regarding their output. The mobility of stars in the industry is needed in order to measure the effect of the hiring of a star. Additionally, there must be some form of development in the industry to enable the
measuring of the effect of developing a star. Law firms are very actively developing their associates in combination with the highest attainable target of becoming a partner with the firm. Finally, law firms do not differentiate between types of work as do other companies such as Shell, Unilever, or ING. This makes it easier to measure differences caused by the performance of hired and internally developed stars. When a company has many different capital streams and activities, it becomes much harder to link the effect of a star to specific performance of a company.
In this methodology chapter firstly, the sample that is used in this research will be presented. Secondly, the dependent variables are presented and the choice for these specific variables will be clarified. Next the independent variables will be presented and how they were constructed will be clarified, followed by the control variables. Lastly, the moderator variables and their construction will be discussed.
Sample and data collection
Law firms are considered to be a good fit to test the effect of hiring and internally developing stars. Data regarding the development of partners and partner movement has been gathered from international law firms in Hong Kong by using the Hong Kong Law Society List and the Hong Kong Bar Association Bar list. The data has been combined and put together in a dataset by Betancourt & Wezel (2016) and made available for use for this thesis.
The dataset covers the complete population of international law firms in Hong Kong over a time frame of eleven years. It thus gives the opportunity to measure very precisely what the effects are of acquiring as well as developing partners on firm performance. The first recording started in January 1998, and the final recording ended in December 2008. During this time period, 987 partners have been recorded of which 250 at one point switched firms (Betancourt & Wezel, 2016). Data has been gathered on both the level of the firm and that of the employees. The sample consists of a total of 9,000 individuals working within 182 different law firms. Approximately 20 percent of these individuals were partners, where 80 percent were associates and support staff. 40 out of the 182 recorded law firms have failed during the eleven-year time period. On the individual level, data has been gathered
concerning the demographic characteristics of the partners. These demographic characteristics are based on the 13 characteristics offered by Martindale-Hubbell who have labelled the different forms of expertise of lawyers (Betancourt & Wezel, 2016). By comparing the
expertise of the partners with the expertise of the firm, the moderator Homogeneous Expertise is created.
Measures
Dependent variables
Both positive and negative consequences were found in the literature review regarding the hiring of stars. Therefore, firm performance will be measured on two dimensions. Stars can have a positive effect on firm performance, however, hiring a star comes at a cost. It is important to measure not only the possible positive effects, but the possible negative effects of stars on performance as well. This will give a more complete picture of the effects of stars on firm performance. Firm growth is interpreted as a positive performance indicator because firm growth represents the fact that a firm is dealing with more work than they can handle, or they expect to be able to handle in the short future. Wiklund (1999) elaborates in his article on performance relationships among entrepreneurs on how firm growth is one of the most
important performance indicators for small firms. Performance has been a subject of interest in the academic literature for many years (Cho & Pucik, 2005) and a variety of indicators have been created and tested. Growth is demonstrated when a firm has been able to increase its size in the past. A firm increasing its size, even at the same profit level, will be able to increase absolute profit and revenue (Santos & Brito, 2012). Growth has been analysed in 88 studies by Capon et al. (1990). They found a consistent relation between growth and higher financial performance. In these studies, growth in assets and sales was shown to have a positive relationship to firm performance. Firms of larger size can create economies of scale and market power, leading to more profitability in the future (Santos & Brito, 2012). In most of the academic research, asset and firm growth has been identified as a positive determinant for firm performance (Cho & Pucik, 2005).
Based on these findings growth has been added as dependent-variable in this thesis. Firm growth is measured on the level of the partners and on the level of partners plus associates. The amount of legal work that requires an x-number of lawyers is represented in
the number of partners and associates working at the firm. This represents the performance of the firm, a firm that hires more partners and associates than rival firms, is outperforming them in terms of the amount of work. Growth has been measured at the level of partners plus associates because this reflects the amount of growth of the value adding aspects of the firm in the best way. Firm growth has been measured on the level of only the partners as well, to specifically test the effect of the partner without possible moderating effects of the associates. This extra test is done to verify if effects that are found in the regression with the partners plus the associates are not heavily influenced by the performance of the associates and to ensure a thorough analysis.
The possible negative consequences of hiring and developing a partner were assessed by measuring firm failure. This happens when firms fail to generate enough income, due to an inability to find enough work. When a firm finds itself in such a situation, an expensive star can drag the firm down. Another possibility is that the expected performance and related pay do not correspond; the star is underperforming and does not deliver the anticipated work. During the eleven years that the law-firms in Honk Kong were observed, 40 law-firms failed. It has been taken into account that the variable Failure is a rudimentary measure as it denotes a simple yes or no for a firm failing. However, this variable does give an insight into the possible negative consequences of law-firms hiring and developing partners. Therefore, by clarifying this extra dimension, it is of value for this research.
Failure
The variable Failure is used by looking at organizational survival in years, in the sense of success or failure, as organizational performance indicator. Failure is coded as a binary variable that starts to record a “1” when a firm has failed in a certain year. Success or failure is often used in studies to measure performance when other data is unavailable (Dess & Robinson, 1984). Each firm is recorded in a timeframe of eleven years. Although some firms
were founded after the start of the collecting of the data, every firm is recorded up to 2008. Therefore, a firm that did not have a recording of 2008, or 2007 etc. is a firm that stopped existing, thus failed. For every firm that did not reach 2008, a score of “1” was given. Using these results, a dummy variable is created specifying a score of “1” for failure and a score of “0” for no failure.
To test for the effect of Internal Development and External Hire on Failure a Probit-regression is used. A Probit-Probit-regression is a nonlinear Probit-regression model that is designed to test the relations of binary dependent variables and binary independent variables. Because a regression with the binary dependent variable Failure tests the probability that Failure notes a score of “1” a nonlinear formulation is required that measures the predicted outcome between “0” and “1” (Stock & Watson, 2015). Therefore, a Probit-regressions is the best fit to test for the effect on the binary Y-variable Failure. When a positive effect is found between one of the independent variables and Failure, it would thus indicate that performance of a firm is negatively influenced by the independent variable. A positive coefficient translates to that an increase in the independent variable means a lower chance of a firm failing. Contrastingly, when a negative effect is found between the independent variable and Failure, it indicates a positive effect on performance. Thus, a negative coefficient will translate to a decrease of the chance of a firm failing when the independent variable increases.
Growth
Growth is constructed using variables from the dataset that denote the number of partners per
firm in a given year and the number of associates in a given year. By using the following formula, the growth in a specific year is calculated:
𝐴𝑚𝑜𝑢𝑛𝑡 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟𝑠𝑡1− 𝐴𝑚𝑜𝑢𝑛𝑡 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟𝑠𝑡0 𝐴𝑚𝑜𝑢𝑛𝑡 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟𝑠𝑡0
Two different growth variables are created. The first growth variable is Growth of Partners. This variable measures the growth of the number of partners per year. The second growth variable Growth Partners plus Associates measures the growth of the number of partners and associates together. The latter has been created to test if firm growth would be affected differently by the dependent variables, Internal Development and External Hire, when the associates were added to the equation. Adding a partner to a firm should be more important to organizational performance than adding an associate; the partner is the individual with the exceptional performance. However, the literature shows that stars rely on others in the
organization to reach their optimum performance (Groysberg, Lee, & Nanda, 2008). This can implicate that firms that grow in the number of partners plus associates perform better than firms that only add partners.
To test for the effect of Internal Development and External Hire on Growth Partners and Growth Partners plus Associates a Linear-regression is used. When the dependent variable is continuous, and the independent variables are binary or dummy variables, a
regression analysis can be used to test the hypothesis, because the mechanics of the regression with binary X-variables works in the same way as it would for continuous X-variables, that require a Linear-regression as well (Stock & Watson, 2015).
When a positive effect is observed between one of the independent variables and
Growth Partners or Growth Partners plus Associates, it indicates that performance of a firm
is positively influenced by the independent variable. Thus, a positive coefficient will translate to an increase in chance of a firm growing when the independent variable increases. On the other hand, when a negative effect is found between the independent variable and Growth
Partners or Growth Partners plus Associates, it would indicate a negative effect on
performance. A negative coefficient will translate to an increase in chance of the firm shrinking when the independent variable increases.
Independent variables
The variable, External Hire YES NO, is operationalized using data from the dataset used in the article of Betancourt & Wezel (2016). External Hire is constructed using five variables from the dataset that recorded an external hiring event on 5 different levels. The variables;
hire_below, hire_partner0, hire_partner1, hire_partner2, and hire_partner3 presented the
number of partners externally hired on that level per year per firm (Betancourt & Wezel, 2016). For every observation of a partner being hired on a certain level per year per firm a score of “1” was given to the variable External Hire. When no such event was observed, a score of “0” was denoted.
The variable Internal Development YES NO is measured using binary variables that tracked the promotion of associates to partner within a firm (Betancourt & Wezel, 2016). The variable Internal developed YES NO was composed of three variables; Difference number of
partners, Number of partners that exit, number of partners externally hired. It enables to
calculate the number of partners per year that must have been internally developed. First, the difference in number of partners per firm per year has been calculated using the following formula:
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟 𝑔𝑟𝑜𝑤𝑡ℎ = 𝐴𝑚𝑜𝑢𝑛𝑡 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟𝑠𝑡1− 𝐴𝑚𝑜𝑢𝑛𝑡 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟𝑠𝑡0
A positive result would indicate the number of partners that were added to the firm. However, a negative score could simply imply that more partners left than were internally developed. For this event the formula has to be compensated in the end. After the difference in number of partners is calculated, the number of partners that have been externally hired should be
subtracted. This then gives the number of partners that have not been hired externally.
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟𝑠 𝑛𝑜𝑡 𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙𝑙𝑦 ℎ𝑖𝑟𝑒𝑑
However, an x-number of partners also leave the firm per year. The result of the formula that calculated the number of partners that have not been hired externally did not compensate for the number of partners that left the firm. Therefore, the result of the formula could be
misleading. Thus, the number of partners that have left the firm should be added to the result of the formula, which leads to the following equation:
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟𝑠 𝑛𝑜𝑡 𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙𝑙𝑦 ℎ𝑖𝑟𝑒𝑑
= 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟 𝑔𝑟𝑜𝑤𝑡ℎ − 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟𝑠 𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙𝑙𝑦 ℎ𝑖𝑟𝑒𝑑 + 𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟𝑠 𝑒𝑥𝑖𝑡𝑖𝑛𝑔 𝑡ℎ𝑒 𝑓𝑖𝑟𝑚
This can be shown with the following example: When a firm goes from 10 partners to 7 partners and simultaneously an exit of 5 partners is observed, it would mean that 2 partners have been added to the firm. When one partner is externally hired the result of the formula would then be 1, as can be seen in the example of the formula below.
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟𝑠 𝑛𝑜𝑡 𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙𝑙𝑦 ℎ𝑖𝑟𝑒𝑑 = −3 − 1 + 5
When no partners have been externally hired, two partners should have been internally developed and the formula subsequently denotes a score of 2.
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑎𝑟𝑡𝑛𝑒𝑟𝑠 𝑛𝑜𝑡 𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙𝑙𝑦 ℎ𝑖𝑟𝑒𝑑 = −3 − 0 + 5
This formula also allows for negative results, meaning that there were more people exiting than developed or hired. A dummy variable was created using the scores above zero to indicate a “1” for an event of internal development and a score of “0” for results of zero or results below zero.
To measure the independent variables, binary variables was created. This implies that only the event of “yes” or “no” of an external hiring of a partner or the internal development
of a partner has been measured. The choice for a binary variable has been made because, the event of an external hire or internal development of a partner gives enough data to test the hypotheses of this research. There is not a large number of observations available regarding more than one partner being externally hired or internally developed, thus numeric values would not provide extra usable information.
Moderator variable
The moderator, expertise, is comprised from data gathered on the individual level of the partners. The international edition of legal directory Martindale-Hubbell offers material on lawyers’ demographic characteristics. The partners have been coded based on the 13 characteristics outlined by Martindale-Hubbell, by Betancourt & Wezel (2016) in their research paper. This data, part of the dataset created by Betancourt & Wezel (2016), on the individual level of the partners is used to create the variables needed to build the
homogeneous expertise moderator. When a firm possessed one or more characteristics of the Martindale-Hubbell scale, it was matched to the individual score of the partner being hired or internally developed to measure expertise overlap. In order to measure for expertise overlap, the variable Homogeneous Expertise was created. When a partner has the same expertise as one of the characteristics of the hiring or developing firm, homogeneous expertise will be given a score of “1”. When the firm does not have overlapping characteristics with the expertise of the partner being hired or internally developed, homogeneous expertise will be rewarded a score of “0”.
To examine the effect of the moderator, the interaction between homogeneous
expertise and the independent variables was calculated. The following interacting variables
were created:
Interaction1:
𝐸𝑥𝑡𝑒𝑟𝑛𝑎𝑙 ℎ𝑖𝑟𝑒 𝑌𝐸𝑆 𝑁𝑂 𝑜𝑓 𝑃𝑎𝑟𝑡𝑛𝑒𝑟 ∗ 𝐻𝑜𝑚𝑜𝑔𝑒𝑛𝑒𝑜𝑢𝑠 𝐸𝑥𝑝𝑒𝑟𝑡𝑖𝑠𝑒
Interacton2:
𝐼𝑛𝑡𝑒𝑟𝑛𝑎𝑙 𝐷𝑒𝑣𝑒𝑙𝑜𝑝𝑚𝑒𝑛𝑡 𝑌𝐸𝑆 𝑁𝑂 𝑜𝑓 𝑃𝑎𝑟𝑡𝑛𝑒𝑟 ∗ 𝐻𝑜𝑚𝑜𝑔𝑒𝑛𝑒𝑜𝑢𝑠 𝐸𝑥𝑝𝑒𝑟𝑡𝑖𝑠𝑒
This process follows the conceptual and statistical model for simple moderation from Hayes (Hayes, 2013). The model for the moderation is presented in figure 2 below.
Figure 2: Moderator and interaction variables model.
Control variables
A number of factors are controlled, as they could influence the effect of stars on organizational performance. Firstly, Firm size can influence the effect of a star on
performance as it will be more significant in a smaller firm (Groysberg, Polzer, Elfenbein, 2011). A count variable will be used for the total number of partners and associates per year.
Homogeneous Expertise
Internal Development Growth of Partners
Internal Development * Homogeneous Expertise Growth of Partners Internal Development Homogeneous Expertise External Hire * Homogeneous Expertise
Secondly, Density will control for the total number of international law-firms in Hong Kong in a specific year. This variable has originally been used to measure how much a firm is able to move in the rankings (Betancourt & Wezel, 2016), but it also helps explaining performance differences between firms in a specific year. To measure this variable, the natural log of the annual gross domestic product of Hong Kong is used. The last factor is Exit of partner. It is important to measure the effect of stars in order to compensate for the effect of stars leaving the firm. The original dataset uses four different binary variables of “0” or “1” to indicate; the exit of a partner to a different firm or market, or the partner leaving Hong Kong (Betancourt & Wezel, 2016). These variables will have to be computed to one single binary variable to indicate a partner leaving.
Results of the Multivariate analysis
The following chapter reports the results of the multivariate analysis performed in this thesis. First, the descriptive statistics of the variables of this research are presented to provide an overview of the data. Secondly, significant correlations are reported based on the performed correlation analysis. Finally, several regressions were carried out to test the hypotheses that are formulated in the theoretical chapter. In this thesis the following hypotheses are tested;
H1: Externally hiring a star has a positive effect on firm performance. H2: Internally developing a star has a positive effect on firm performance.
H3: Internally developing a star has a more positive effect on firm performance than externally hiring a star.
H4: Homogeneous expertise of a partner will decrease the positive effect of externally hired stars on firm performance.
Descriptive statistics and correlation analysis
Table 1 reports the descriptive statistics and the correlation analysis of the variables of the used dataset. The relationships were tested using the Pearson product-moment correlation coefficient (Pearson, 1896). First, the correlations between the dependent variables and control variables are presented. Secondly, the correlations between the dependent variables and the independent variables are shown. Finally, the correlations between the moderating variables and the independent variables, and the correlations between the moderating variables and the control variables are outlined.
While studying the correlations between the dependent variables and the control variables, failure was shown to have a very weak non-significant negative correlating effect with Firm size (r=-0.055), and density (r=0.013). Furthermore, Failure has a weak significant correlating effect with Exit Partners Total (r=0.124). The independent variable Growth
Partners showed a significant correlation with Firm size (r=0.198), the effect is weak.
Moreover, Growth Partners has a very weak negative correlating effect with Density
(r=-0.055), and also a very weak correlating effect with Exit Partners Total (r=0.062). The final
independent variable, Growth Partners plus Associates, exhibited a weak correlating significant effect on the control variable Firm size (r=0.187). Additionally, a very weak negative effect can be observed on density, and a very weak positive effect on Exit Partners
Total (r=0.032).
Regarding the correlation between the independent variables and the dependent variables the following relations were observed. The independent variable Internal
Development is negatively and very weakly correlated with Failure (r=-0.035). A similar
relation can be observed regarding the independent variable Hire External, as it shows a very weak negative relation with Failure as well (r=-0.013). Internal Development has a
relation with Growth Partners plus Associates (r=0.383). Contrastingly, Hire External shows much weaker non-significant relations with Growth Partners (r=0.058) and Growth Partners
plus Associates (r=0.028).
Looking at the relations between the moderating variables and the independent
variables, the following can be observed. Internal Development has a non-significant and very weak negative relation with Interaction 1 (r=-0.010). However, Internal Development has a significant and moderate relation with Interaction 2 (r=0.494). The precisely opposite effect can be observed in the relations between Hire External and the moderators. Hire External has a significant and positive relation with Interaction 1 (r=0.407), and a non-significant and very weak negative relation with Interaction 2 (r=-0.012).
Finally, the relations between the moderating variables and the control variables are discussed. Interaction 1 has a non-significant and very weak and negative relation with both
Density (r—0.032) and with Exit Partners Total (r=-0.015). Interaction 1 has a very weak
and non-significant positive relation with Firm size (r=0.070). Interaction 2 has a significant and positive relation with both Firm size (r=0.095) and Exit Partners Total (r=0.134). The relationship between Interaction 2 and Density is very weak negative and non-significant
Table 1: Descriptive statistics.
Mean St. Dev. Min Max 1 2 3 4 5 6 7 8 9 10 11
1 Failure 0,021 0,142 0 1 1
2 Growth Partners 0,159 1,272 -1 17 -0,043 1 3 Growth Partners plus
Associates
0,569 3,326 -1 64 -0,054 0,624** 1
4 Firm size 17,082 23,883 0 172 -0,055 0,198** 0,187** 1
5 Density 131,598 10,913 119 154 0,013 -0,055 -0,039 0,005 1
6 Exit Partners Total 0,281 0,593 0 5 0,124** 0,062 0,032 0,141** -0,135** 1 7 Internal Development 0,376 0,485 0 1 -0,035 0,513** 0,383** 0,125** -0,062 0,225** 1 8 Hire external 0,091 0,288 0 1 -0,013 0,058 0,028 0,068 0,001 -0,002 0,105** 1 9 Homogeneous Expertise 0,023 0,230 0 6 -0,013 0,123** 0,083* 0,113** -0,034 0,051 0,345** 0,051 1 10 Interaction 1 0,001 0,033 0 1 -0,005 0,005 -0,003 0,070 -0,032 -0,015 -0,010 0,407** 0,143** 1 11 Interaction 2 0,489 0,500 0 1 -0,017 0,226** 0,177** 0,095* -0,045 0,134** 0,494** -0,012 0,700** -0,005 1 **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-taile
Regression analysis
For hypothesis 1 the effect of externally hiring a partner on firm performance is tested. For hypothesis 2 the effect of internally developing a partner on firm performance is examined. Firm performance has been measured on multiple levels. This thesis used three dependent variables to measure firm performance. Firm performance was measured with the help of the variables Failure, Growth Partners, and Growth Partners plus Associates. The control variable Density was used to control for the total amount of law-firms and helped explaining differences in firm size in a specific year. It controls for possible differences in the size of the market and thus amount of work per year.
Table 2 shows the results of the Probit-regression performed to test the relations with
Failure. Table 3 and table 4 show the results of the linear regressions that examined the effect
of the independent variables Internal development & Hire External on the dependent
variables Growth Partners, & Growth Partners plus Associates. In each of the tables, model 1 represents the base model comprised from the control variables, model 2 represents the base model with addition of the independent variables, and model 3 displays the base model with addition of the independent variables and the moderator variable plus the effect of the interacting variables.
Firstly, the model with the results from the analysis on Failure will be discussed. This research started with testing the effect of the non-lagged variables. However, the results showed only omitted variables. Due to the low amount of failures and the fact that binary variables are used for both Failure and the independent variables, the chance of getting a score of “0” at the identical observation as Failure is large. This is problematic for measuring the effect of the independent variables and the moderator on Failure. However, the effect of a star dragging a firm down due to low performance is usually not immediately visible. Firms that are financially healthy enough to acquire or develop a new star will not immediately go
out of business when a star is underperforming. Lagged variables were created that measured the effect of hiring and internally developing partners on the chance of a firm failing for up to five years. Significant results were found.
The analysis of the results proved that Interaction 1 was a perfect predictor and was excluded from the final regression. If Interaction 1 was to be included in the regression, it would have a very significant result for every lagged year, since there is no variance. Consequently, because there is no variance, there is no way to determine whether the significant effect of Interaction 1 is correct or accurate as the relationship between the variables Interaction 1 and Failure is a constant in the observed data.
Table 2: Summary of Hierarchical Probit-Regression Analysis for Variables Predicting Failure.
Note: *,**,*** denote a rejection of the null hypothesis at the 1,5,10 percent significance level respectively. The
variables have been lagged for 5 years.
Model 1 Model 2 Model 3
Variable Control SE B Main SE B Interaction SE B
Firm size -0.042* 0.012 -0.469* 0.013 -0.049* 0.014
Density -0.003 0.007 0.002 0.008 0.001 0.008
Exit partners total 0.378* 0.109 0.395* 0.110 0.409* 0.113
Internal development Lag 0.024 0.250 -0.029 0.303
Hire external Lag 0.799** 0.370 0.834** 0.405
Homo expertise Lag -0.889 0.313
HomoExp. x Int.Dev. Lag 1.565 0.214
Prob. Chi2 0.000 0.000 0.000
LR Chi2 27.12 31.05 33.25
When looking at model 3, the first aspect that stands out is that Firm size has a small negative and significant effect of 5 percent on Failure. This implies that bigger firms have a smaller chance of failing. Secondly, Exit Partners Total has a positive and significant effect of 41 percent on a firm failing within five years. This means that when stars leave a firm, the firm has a bigger chance of failing. When looking at the effect of the independent variables in model 3, a significant and positive effect was observed for externally hiring partners. This result implies, that when a firm externally hires a partner, the chance of failing within 5 years increases with more than 80 percent. No significant effects were observed between the moderator and interaction variable Internal Development x Homogeneous Expertise Lag in model 3. Therefore, it cannot be proven that the homogeneous expertise of a partner influences the effect of internal development on firms failing.
The following effects were found in the different models. Firms size in model 1 is very similar to model 3, however, a jump can be denoted in model 2. Without the moderator,
Firm size has a bigger influence on firms failing. Furthermore, adding the moderating variable
to the analysis does not have a large influence on explaining the model. The results from the
LR Chi2 test for the probability of minimally one of the predictors effect being not equal to zero. The Probability Chi2 presents scores of 0.000 for each model, indicating that the results from the LR Chi2 are significant and that at least one of the coefficients is not equal to zero.
Table 3 presents the results of the regression that measured the independent-variables against the dependent-variable growth of partners. See table 3. The theory regarding star performers has shed light on multiple positive and negative associations with the hiring and developing of stars. The effect of a star’s performance could drop for up to five years before the star is back at its original level (Groysberg, Lee, & Nanda, 2008). Therefore, the
appearance of positive effects of the acquirement of a partner could be delayed. To test this assumption lagged variables were created for internal development and external hire of
partners. The internal development and the external hire of partners were lagged up to five years to examine the different effects of the X’s on the Y-variables. No significant changes were found for external hire and internal development, that were lagged up to 5 years, on
Growth of Partners and Growth Partners plus Associates.
Table 3: Summary of Hierarchical Regression Analysis for Variables Predicting Growth of Partners.
Note: *,**,*** denote a rejection of the null hypothesis at the 1,5,10 percent significance level respectively. A negative and significant coefficient of -0.13 was found for the variable Exit
Partners Total in model 3. This indicates that when one partner exits a firm, the growth of the
number of partners of the firm lowers with 13 percent. Internal development of a partners has a significant effect on the firm growth at the level of the partners. Internal Development presents a coefficient of 1.53 which translates to a growth of 153 percent of Growth Partners when a partner is internally developed. Because this research uses binary variables that are only interested in the event of a partner being internally developed or hired externally, the
Model 1 Model 2 Model 3
Variable Control SE B Main SE B Interaction SE B
Firm size 0.009* 0.002 0.006* 0.001 0.005* 0.001
Density -0.006*** 0.004 -0.005 0.003 -0.002 0.002
Exit partners total -0.028 0.062 -0.353* 0.058 -0.129** 0.054
Internal development 1.298* 0.073 1.533* 0.111 Hire external -0.056 0.112 -0.067 0.271 Homo expertise -0.248** 0.126 HomoExp. x Ext.Hire 0.270 0.667 HomoExp. x Int.Dev. 0.131 0.278 R2 .037 .249 .293 F 14.45 74.17 34.51 N 1126 1126 677
number of partners being added is not counted. The growth of 153 percent could imply that when a firm internally develops a partner, they internally develop more than just one. When examining the effect of the moderator and interacting variables, a significant main effect can be denoted of -0.248 (P<0.05). The effect of Homogeneous Expertise is present but not conditional on the value of Internal development or External hire of partners
The results of the different models show no effect on Firm size when adding the independent variables and the moderating variables. For each of the models 1, 2, and 3, Firm
size shows minimal change. When the independent variables were added in model 2, an
increase of the effect of Exit Partners Total on the dependent variable became visible. However, this effect is noticeably smaller in model 3, when the moderating variables were added. It thus seems that the interacting variables lower the negative growth effect of partners leaving the firm. The opposite reaction is found for the dependent variable Internal
Development. The effect of Internal Development on the dependent variable Growth
Partners, actually became stronger, when in the interaction variables were added in model 3.
An increase of the coefficients from 1.3 to 1.5 is observed. Internally developing
homogeneous partners within firms seems to have a stronger effect on the growth of the number of partners than non-homogeneous partners. This could be explained by the fact the firms that internally develop partners transfer knowledge from within the firm to the partner, which leads to the generating of more of the same expertise.
30 Percent of the model is explained by the variables used in the models. A large jump of 0.212 in the R2 from model 1 to 2 was observed and the R2 showed a much smaller change when the interaction variables were added. Especially the adding of the independent variables seems to be important to explain the effect of the model.
The final regression table, table 4, shows the results of the effects of the independent variables on the dependent variable Growth Partners plus Associates. See table 4.
Table 4: Summary of Hierarchical Regression Analysis for Variables Predicting Growth of Partners plus associates.
Note: *,**,*** denote a rejection of the null hypothesis at the 1,5,10 percent significance level respectively. Model 3 of the regression shows significant effects for internal developed partner. The internal development of a partner has a very strong effect on the growth of the firm regarding partners and associates. When one partner is internally developed, firms are likely to grow with 200 percent. A similar but smaller effect is observed for the effect of internally
developing a partner on Growth of Partners. An explanation for the larger effect could be the adding of the associates to the equation. When partners are internally developed, more
associates are hired as well. The variable Firm size has a significant effect on the growth of the firm’s partners and associates. Although the effect is small compared to the effect of the internal development of stars. When a firm grows with one, growth of partners and associates is likely to raise with 0.8 percent. This can indicate that the overall growth of the firm has a small effect on hiring or developing more partners and associates. Exit Partners Total has a
Model 1 Model 2 Model 3
Variable Control SE B Main SE B Interaction SE B
Firm size 0.023* 0.004 0.018* 0.004 0.008* 0.002
Density -0.012 0.009 -0.011 0.009 -0.004 0.004
Exit partners total -0.149 0.157 -0.591* 0.159 -0.287* 0.101
Internal development 1.844* 0.199 2.006* 0.201 Hire external 0.162 0.310 -0.399 0.512 Homo expertise -0.549** 0.236 HomoExp. x Ext.Hire 0.559 1.264 HomoExp. x Int.Dev. 0.869*** 0.492 R2 .030 .0933 .190 F 12.56 25.31 21.20 N 1236 1236 733
strong negative and significant effect on Growth Partners plus Associates. The coefficient presents a score of -0.287 which means that a partner exiting the firm lowers the growth of the firm’s partners and associates by almost 30 percent. This result implies that when stars leave the firm, this lowers the performance of the firm.
When studying the effects of the interacting variables, a significant effect on a scale of the 10 percent significance level was found for Interaction 2, Internal Development x
Homogeneous Expertise. The coefficient denotes a positive effect of 0.87. First, it is
important to consider that both the independent and the moderator variables are binary. Low internal development is interpreted as no internal development, whereas high internal development is interpreted as the event of a firm internally developing a partner. Secondly, when a star’s expertise scores a match with the firm, Homogeneous Expertise scores a “1”. When a star’s expertise does not score a match with the internally developing firm this event is interpreted as heterogeneous development and is given a score of “0”. When studying the interaction graph, see figure 3, the following interaction effect can be observed.
Firms that score low on the development of partners tend to grow on partners with a heterogeneous expertise. When firms score high on the internal development of partners, firms tend to develop more partners with a homogeneous expertise. This result implies, that firms that are internally developing partners duplicate the existing knowledge present in the firm. However, homogeneous expertise shows a marginal stronger effect on growth than heterogeneous expertise. Because the significance of the effect is marginal and only shows a significance on the 10 percent level, the results are not conclusive. The expertise of a star does not influence the relation between performance and the hiring or internal development of a partner.
Finally, the differences between the models were studied. The complete model
0.09 when the independent variables are added and changes from 0.09 to 0.19 when the interacting variables are added in model 3. Again, this indicates that adding the independent and interacting variables has a big effect on explaining the variance of the entire model.
Figure 3: The moderating effect of homogeneous expertise on the relationship between internal development of partners and organizational performance.
Discussion
This thesis wanted to test if externally hiring or internally developing stars may influence the effect of stars on organizational performance. Furthermore, the moderating effect of a
homogeneous expertise of a star was tested. The results of the Probit and Linear regression analysis partially support the hypothesis. However, the hypothesized effect of the interacting variables could not be supported. The following section will discuss the research of this thesis and its findings. First, the major findings of this thesis are discussed; the hypothesized
relations will either be supported or rejected, and further finding will be elaborated on. Secondly, the contributions of this thesis will be discussed. Finally, the limitations of this research are presented together with recommendations for future research.
Major findings
There is substantial debate on the effectiveness of the acquirement of stars on organizational performance. Some studies present positive effects of the acquirement of stars on
organizational performance. They state how stars can help attracting new talent and resources and boost the development of knowledge and new skills (Groysberg, Polzer, Elfenbein, 2011; Olroyd, & Morris, 2012; Grioriou, & Rothaermel, 2014; Agarwal, McHale, & Oettl, 2013; Zucker, & Darby, 1997). However, other studies present also negative consequences that come with the acquirement of stars. The portability of a star’s exceptional performance can be less than originally expected. Furthermore, stars can show competitive behaviours that lead to a decrease in group performance. Finally, acquired stars can experience uncooperative
behaviour from resentful managers and other employees whilst trying to learn the new procedures at the organization (Groysberg, Polzer, Elfenbein, 2011; Overbeck et al., 2005; Tiedens, & Fragale, 2003; Groysberg, Nanda, & Nohria, 2004; Kehoe, & Tzabbar, 2014; Hambrick, 1994). Based on the studied literature, the following hypotheses were tested.
H1: Externally hiring a star has a positive effect on firm performance. This hypothesis
is not supported by the results. The performed regressions do not show any significant results for either positive or negative effect of externally hiring a partner on Growth of Partners, or
Growth of Partners plus Associates. Contrarily, a positive and significant effect has been
found regarding the relation between externally hiring a partner and firms failing. It was found that firms that externally hire a partner have an 80 percent increased chance of failing within 5 years. Therefore, when firms externally hire a partner, firm failure is more likely.
This result could be explained by the findings of Groysberg, Lee, & Nanda (20008), who discussed the limited portability of a star’s exceptional performance. Moreover, it could very well be possible that firms overpay the newly hired star whilst the star is struggling to
reach its former performance due to unfamiliar organizational systems and a lacking network (Bidwell, 2011).
H2: Internally developing a star has a positive effect on firm performance. This
hypothesis is partially supported by the results. There is no significant and negative effect observed for Failure. A negative coefficient of Internal Development on Failure would suggest that internally developing partners lowers the chance of failing. However, there are very positive and significant results found for the effect of Internal Development on the performance indicators Growth Partners 1.53 (P<0.001) and Growth of Partners plus
Associates 2.01 (P<0.001). This suggests that when one partner is internally developed, firm
growth of partners and the firm growth of partners plus associates rise respectively with 153 and 201 percent.
H3: Internally developing a star has a more positive effect on firm performance than externally hiring a star. Where there is no significant effect found of External Hire on any of
the dependent variables that acted as performance indicator, there is a strong significant effect found for Internal Development. Therefore, this hypothesis is supported. Internally
developing a star is more likely to have a positive effect on organizational performance than externally hiring one. The strong results found for the internal development of a partner also support the findings discussed by Bidwell (2011), who found that externally hired employees perform less than internally developed employees. Firms have more information on their own employees and can therefor better assess their qualities, lowering the chance of overpaying. Secondly, internally developed employees have better knowledge of the organization and its processes which leads to higher performance compared to externally hired employees (Bidwell, 2011). Furthermore, this result could be explained by the fact that internally developed employees have less chance of uncooperative behaviours from colleagues over resentful feelings (Groysberg, Nanda, & Nohria, 2014). All in all, it seems that firms enjoy
better performance from their internally developed stars because internally develop stars are more familiar with their organization and because it is a more harmonious process than the acquirement of a star.
H4: Homogeneous expertise of a partner will decrease the positive effect of externally hired stars on firm performance. The results of the performed analysis show no support for a
relation between homogeneous expertise and the acquiring of partners. Therefore, the hypothesis is not supported.
H5: Homogeneous expertise of a partner will decrease the positive effect of internally developed stars on firm performance. This hypothesis is not supported. Only the results of the
regression that tested the effect of the interaction variable on Growth of Partners plus
Associates showed a significant effect of 0.869 (P<0.1). The moderating effect of a star’s
expertise on the internal development of a partner presents the opposite moderating effect to what was hypothesized. When there is more homogeneous expertise added to a firm, the positive effect of internal development of partners on growth becomes even larger. This result could be explained by further examining the knowledge available within a firm. When firms internally develop their associates to the level of partnership, they do so by using existing knowledge from within the firm. This would indicate that the expertise of the developed partner is similar to the expertise of the firm. However, the results do not give a very conclusive answer, as the relation is only significant at the 10 percent level. Moreover, the results of the moderating effect of a star’s expertise on acquiring or developing a star mostly show no effect. It can thus very well be possible that the expertise of a partner does not matter for the hiring or developing in a firm. In the end, good partners bring in a large number of clients and it is not unthinkable that this is much more important for the firm than the expertise of the new partner.
