MASTER’S THESIS
The role of CEO demography and organization
characteristics in explaining firm performance:
An Empirical Study of Private Firms in Vietnam
PHAN ANH TU
MSc in International Economics and Business Faculty of Economics
University of Groningen
Supervisor 1: Dr GJALT DE JONG
Supervisor 2: Dr GER LANJOUW
The role of CEO demography and organization
characteristics in explaining firm performance:
An Empirical Study of Private Firms in Vietnam
ABSTRACT
1This study investigates the various impacts on firm performance of CEO demography and firm characteristics with regard to an industrial environment, within a sample of 606 private firms from six provinces in the Mekong River Delta (MRD) in Vietnam. Controlling for firm size, level of competition, industrial sector, and ownership types, the study found that individual CEO demographic characteristics and firm characteristics are directly related to firm performance. Specifically, the results suggest an inverted U-shaped relationship between CEO tenure and firm performance. While CEOs with higher education and higher levels of informal education are strongly associated with higher firm performance, CEO risk-averse behaviour has no connection. According to firm traits, firm age shows a negatively significant association with firm performance. While we found a curvilinear relationship between bribery and firm outcomes, employees’ education has a positive significant link to firm performance. The study reinforces the importance of CEOs in private firms in the MRD, compares the results with earlier international studies and suggests further research in this context.
Keywords: CEO tenure, CEO formal and informal education, CEO risk aversion, bribery, private firms.
1
CONTENTS
---***---1. INTRODUCTION ... 5
2. THEORETICAL BACKGROUND AND HYPOTHESES ... 10
2.1 Upper-echelon theory ... 10
2.2 Contingency properties ... 12
2.3 Hypotheses ... 16
2.4 The theoretical model ... 21
3. RESEARCH METHODS... 22
3.1 Data collection and sample... 22
3.2 Measures ... 24
3.3 Analysis ... 26
4. EMPIRICAL RESULTS... 27
4.1 Sample characteristics ... 27
4.2 Description... 31
4.3 Regression results ... 34
4.4 Robustness analysis ... 37
5. CONCLUSIONS ... 38
5.1 Conclusions ... 38
5.2 Limitations and future research ... 43
REFERENCES... 44
LIST OF TABLES AND FIGURES
Table 1: Key indicators for Vietnam ... 51
Table 2: Number of firms in Vietnam... 51
Table 3: Profit/sale − industrial sectors... 52
Table 4: List of independent variables ... 52
Table 5: Sample distribution − industry sectors and provinces ... 54
Table 6: Descriptive statistics ... 32
Table 7: Correlation matrix... 33
Table 8: Regression result (model 1)
...
36Table 9: Regression result (model 2) ... 55
Table 10: Summary of empirical findings ... 41
--***-- Figure 1: Normal standard distribution ... 53
Figure 2: Sales and ownership ... 28
Figure 3: CEO tenure ... 29
Figure 4: CEO education... 29
Figure 5: Ownership types ... 30
Figure 6: Firm age ... 30
LIST OF ABBREVIATIONS
BLUE
Best linear unbiased estimator
CEO
Chief Executive Officer
MRD
Mekong River Delta
OLS
Ordinary least squares
ROI
Return on investments
R&D
Research and Development
SOE
State-owned enterprise
TMT
Top management teams
VND
Vietnamese dong
1. INTRODUCTION
In recent years, Vietnam has achieved a high growth rate in Gross Domestic Product (GDP), with an average growth rate of more than seven percent during the period from 2001−2005. GDP is expected to increase significantly in the coming years, as Vietnam became an official member of the World Trade Organization (WTO) in 2006. Obtaining this high economic growth rate has involved a considerable contribution from private firms. However, while the economic environment has allowed the number of private firms to increase significantly, there are large differences in firm performance with regard to profit. See Table 1 for an overview of GDP growth rate and Table 2 for an overview of the profits of firms.
[Insert TABLE 1, 2 about here]
More opportunities have been created by a stimulatory investment policy, the enterprise law promulgated in 2000, a new Constitutional Law in 2001, and the Decree 90/2001/CP in 2001. Also, transparent laws and a more open business environment have, in turn, stimulated the establishment of many new firms. However, an increase in the number of new firms created does not guarantee the quality of growth. For instance, faced with a small business scale and many constraints, including limited access to loans, outdated technology and low education levels among CEOs (Ninh et al., 2007), firm performance improvement related to investment expansion capability is limited. Unsurprisingly, Vietnam − a country in transition from a centrally planned economy to a market economy − is attempting to integrate into the world economy, and as a result firms are constantly attempting to improve their performance consistent with this process of integration in order to deal with the fierce competition in the international arena. Chief executive officers (CEOs) and employees of firms, in fact, have started to professionalize in line with the global dynamic business environment. Therefore, in Vietnam the roles of managers and the human capital strategies of firms are extremely important in relation to reaching important economic targets.
relatively little influence on firm outcomes due to environmental and inertial forces. Using a sample of 167 companies over a twenty-year period, they statistically analysed the proportion of company performance (as measured by sales, profits, and return on sales) that could be attributed to the top executive of a firm in a given year. They concluded that leadership only accounted for between 6.5 and 14.5 percent of variance in the three performance measures tested. Astley and Van de Ven (1983) revealed that the managerial role can even be described as inactive with respect to organizational outcomes. Although concern about the role of top executives is not new, some recent studies have gone beyond earlier prescriptions for effective management. These new approaches concentrate especially on the key role of background, tenure, experience, and other values of top managers in explaining the choices they make related to organizational outcomes. Hambrick and Mason (1984) argued that both firm performance and strategic choices are linked with the characteristics that top executives display within a firm. Karami et al. (2006) argued that a high level of management quality, measured in terms of managerial characteristics such as capabilities, skills, expertise, age and education, is more common among successful than unsuccessful firms.
Many researchers have found direct effects of the top management team’s (TMT) demography on firm performance (Pleffer, 1983; Murray, 1989; Eisenhardt and Schoonhoven, 1990; Ken G. Smith et al., 1994). Ninh et al. (2007) found that the ability of a firm to achieve investment expansion depends on the firm’s ability to access loans for investment expansion, past sales growth and profit, and CEO educational level, all of which may lead to an indirect increase in sales growth. However, the researchers did not show whether CEO demographic characteristics directly impact on firm performance. While past research has examined the relationship between a set of organizational antecedents of succession outcomes (firm growth, profitability, R&D intensity) and the characteristics of CEOs selected by both insiders (Dalton and Kesner, 1983) and outsiders (Datta and Guthrie, 1994), the set of organizational and executive characteristics studied has been rather limited. Therefore, we argue that in the current literature too little attention has been paid to the CEO alone in relation to the issue of how firm characteristics directly impact on firm performance. Furthermore, much of the literature on CEO demographic variables and TMT has been developed in Western countries, and it is interesting to assess whether these theories are applicable in the case of Vietnam.
firms, which include various types of ownership, for example, private companies (sole proprietorship) and households (family-owned businesses), joint-stock companies, limited-liability firms and collectives (cooperatives). Interestingly, while the number of new private firms increased significantly, their performance differed across all industrial sectors (Ninh et al., 2007). Specifically, it has been asserted that the average growth rate in the number of private firms in the period 2000−2004 was around 10 percent compared to 5.3 percent in the period 1986−1990. Surprisingly, the profit-to-sales ratios of both trading (36.1 %) and service industries (33.4%) are far above that of the manufacturing industry (7.7%). This huge gap can primarily be explained by the operating costs. In addition, an increase in the number of private firms was the indirect result of the transformation of state-owned enterprises (SOEs) into joint-stock companies which began in 1992. The main reasons for this reform were to restructure SOEs and encourage employees to work efficiently (Loc, 2006). Another issue to consider is that SOEs still play an important role in many key economic sectors in the Vietnamese economy, and they receive more capital investment from the government than private firms but do not perform well and have even shown significant decreases in profit ratio (see Table 2). While private firms have recently become appreciated in Vietnam’s economic development strategy as they contribute to reduced unemployment, these firms face many constraints, such as an excessive level of administration, the need for loans, a lack of opportunity to access loans, small business scale, a lack of innovation, outdated technology, and the low level of education of managers.2
In short, the aim of this study is to explain the differing performances of firms in Vietnam. More specifically, we will focus on private firms in Vietnam’s MRD. The reason why we have selected the MRD region as a central area for our study is that the number of private firms found here account for one-quarter of the total number of private firms within Vietnam. While these firms play an important role in creating jobs and income for residents, they contribute less than 10 percent of the GDP of the MRD region (Taussig, 2003, p. 4). Private and household (family-owned) businesses predominate, characteristically operating on a small scale due to the fact that the effects of purely professional agricultural production and a risk-averse psychological attitude leads to firms specializing in the business sector in order to rapidly retrieve the initial capital outlay (Ninh et al., 2007). Additionally, the role of private firms in the
2
MRD is becoming increasingly important in contributing to the economic development of the region, contributing especially to the total value of industrial production (see Table 3, appendix), while foreign direct investment (FDI) has started to decrease and has not yet provided a strong stimulus for economic development or the transformation of the economic structure within this region. We will not suggest reasons for the decline in FDI, instead we will aim to investigate the determinants of private firm performance.
By looking at different patterns of internal and external effects on private firms’ performance in the MRD, the following research questions will be answered: To what extent is the difference in firm performance affected by the CEO’s personal characteristics? What is the relationship between CEO tenure, CEO education (either formal or informal), risk-averse behaviour and firm performance? To what extent is the difference in firm performance affected by firm characteristics? What is the relationship between firm age and firm performance? Is there a relationship between firm performance and the amount of money used for bribes? Is there any connection between employees’ education and organizational outcomes?
This thesis contributes to the study of the effects of CEO demography and firm characteristics on firm performance in the following ways. First, this study treats equally the quantitative effects of individual CEO demographics and organizational characteristics as well as the industry environment on firm performance of private firms in the MRD. Amongst the managerial characteristics examined are CEO tenure and CEO educational level for both formal education and informal education,3 as a new perspective added to upper-echelon theory and human capital literature in our analysis. Many prior studies have indicated that the CEO’s age is a predictor of differences between firms in terms of performance in general and developing business strategies in particular. However, we will not focus on this point; instead, we will concentrate on whether CEO risk-averse behaviour affects firm performance.
Second, while past research has extensively examined the relationships between organizational factors and CEO characteristics on firm performance (Dalton and Kesner, 1983), little attention has been paid to the environment, industry and the specific organizational context within which CEOs operate. We will consider firm characteristics such as firm age, employees’ educational level and bribery as new and additional characteristics contributing to
3
organization research theory and having a direct or indirect impact on firm performance. At the same time, we treat some factors related to the industry environment as well as organization characteristics as control variables in the research model, for example, level of competition, industrial sectors, ownership types and firm size (number of employees).
In the analysis of the relationships between CEO demographic variables and organization characteristics and their effect on firm performance, this study presents unique data. In contrast to earlier research, this study is based on both primary data collected from CEOs (or the owners) directly responsible for their firms as well as on secondary data. The sample included many different types of firms, varying in business scale, industry sector and ownership. The data were obtained by means of extensive direct interviews of the managers, with the sample including 606 private firms in 12 provinces in the MRD during the period 2004−2005. This enabled us to develop a good insight into both factual information and subjective interpretations. In particular, through the use of statistical software packages (SPSS, Eviews, Excel), we introduced an econometric model (OLS regression model) in order to assess the impacts of CEO demography and firm characteristics on firm performance. Previous studies in the area are also discussed to support our argument. Therefore, this study provides an extremely important basis for NSE development strategy, not only for all the provinces located within the MRD but for the whole country. More importantly, on the basis of our results local authorities can efficiently modify the business environment and even adjust policy more effectively in support of their aim of stimulating investment from the non state-owned sector, and with the expectation of achieving a higher economic growth rate which is a part of the targets of the modernization and industrialization programme in Vietnam.
2. THEORETICAL BACKGROUND AND HYPOTHESES
The theoretical model of this study is primarily derived from theories that explain the importance of CEO behaviour and characteristics. However, the impact of CEO behaviour on firm performance is contingent on characteristics of the firm and the external environment of the firm. In what follows, we will first summarize the theoretical background of this study. We will then formulate the hypotheses and conclude with the specification of the theoretical model.
2.1 Upper-echelon theory
The upper-echelon theory concerning CEO demographic characteristics as set forth by Hambrick and Mason (1984) has had a major influence in explaining organizational outcomes. According to their view, the spectrum of capital variables believed to impact on organizational outcomes is extended with observable demographic characteristics such as tenure, functional (types of education) and occupational background, and unobservable psychological characteristics (for example, managers’ values and personal characteristics refer to certain other values, attitudes and frames of reference which influence strategic decisions and, thus, firm performance). Consistent with this theory, CEO demography in our study is also defined in terms of both observable (tenure, formal and informal education) and unobservable personal attributes (CEO risk aversion). Hambrick and Mason (1984) claim that organizational performance reflects the top executives’ personal characteristics and that upper-level managers have an important impact on organizational outcomes because of the decisions they are empowered to make for the firm. Hrebiniak and Joyce (1985) proved that through their choices managers greatly affect outcomes. Finkelstein (1992) suggested that the upper-echelon perspective should be expanded to take into account how managerial power affects the association between top managers and organizational outcomes. He pointed out how power may emanate from a manager’s personality. In a similar manner, Hambrick and Finkelstein (1987) suggested that the role of top executives’ managerial power depends on whether discretion is low or high and that the upper-echelon theory will respectively have weak or strong explanatory power.
that firm performance will be more variable as decision-making power becomes more centralized in the hands of the CEOs. This effect will be stronger when one of the firm’s founders is the CEO, a situation which seems to have the greatest effect on firm performance. Obviously, these managers make decisions consistent with their cognitive base, which is in part a function of their personal values and experiences. These personal experiences and values can therefore be associated with organizational outcomes and hence are reflected in their strategic choices (Finkelstein and Hambrick, 1996). They argue that the behavioural view of decision-making is relevant when considering top managers who face great complexity and ambiguity in their tasks. They must select appropriate information and choose how to respond. Defining what is important depends on their interpretation of the situation. Although the effects of education level on organization performance are not as widely observed or clear cut, one can conceive of growth as a performance indicator, in which case executive education levels seem to have a salutary effect (Norburn and Birley, 1988).
Wiersema and Bantel (1992) found that attributes such as organizational tenure, education level and functional backgrounds were useful indicators of an organization’s propensity for strategic change. It has also been found that firm performance is not solely influenced by the education level of the CEO. Other important features include senior managers with managerial experience,4 who tend to place more weight on the process of developing formal strategies than those who lack the relevant managerial work experience (Karami et al., 2006). A prior study has determined that there was a relationship between financial performance and the average and spread of firm tenure, diversity in age and top level experience (Bosman et al., 1997). Pleffer (1983) found the basic rationale for expecting direct relationships between TMT demography and organizational performance while Ken G. Smith et al. (1994) found TMT demography both indirectly related to firm performance through processes,5 and processes directly related to performance. Several empirical studies have provided additional evidence, assessing the direct or indirect impact of top executives’ personal attributes on firm performance that, based on the human capital theory, focused mainly on survival chances (Bosman et al., 1997). However, others have failed to find a significant relationship between TMT demographics and performance (West and Schwenk, 1996).
4
Managerial experience is the total experience in management regardless of the industry (Lee and Tsang, 2001).
5
Similarly, Finkelstein (1988) found that TMT characteristics are a better predictor of organizational outcomes than the CEO features alone, while there is very little empirical evidence on whether the qualities of the CEO alone or those of a broader team are a better predictor of organizational outcomes. Nevertheless, this consistently supports the conclusion that the TMT has greater effect (Hage and Dewar, 1973).
In conclusion, the important feature of Hambrick and Mason’s upper-echelon perspective is their primary focus on the notion of a TMT team rather than the CEO alone. TMT theories are important in all cases but may not be so for the case of Vietnam. A special feature of our research that makes it different to prior studies is that it shows that CEOs in private firms who have complete decision-making power are often the founders of the firm. As a result, CEOs who believe that their destiny lies in their own hands are more likely to try to control firm performance actively (Miller et al., 1982). Consequently, traits of CEOs alone, organizational features, and the differences among them, all of which are associated with firm performance, have emerged as topics of considerable interest in both the academic and popular business literature on executive leadership and organization theory. More crucially, the concentration on CEOs in research on executive leadership appears to be based on the assumption that CEOs have a significant impact on both organizational activities and performance (Finkelstein, 1998).
2.2 Contingency properties
The vast majority of recent research focuses on the relationships between executive characteristics and organizational performance. In some cases these are direct, more often they are contingent on the association between executive characteristics and performance, depending on specific organizational conditions (Finkelstein and Hambrick, 1996). Population ecologists especially have claimed that organizations are hemmed in by environmental and organizational constraints. In this section we will briefly review the key aspects of the theoretical debate on contingency insofar as it relates to the expected association between organization and industry environment characteristics with respect to their effect on firm performance.
Economists such as Adam Smith viewed land, labour and capital as the three production factors that primarily contributed to wealth. Karl Marx’s labour theory held that workers produced all value. Neoclassical theory posited a labour market that was a ‘price-auction’ mechanism where ‘atomistic workers’ offered their services to competitive employers for a wage. Contemporary workforce theories, however, do not take into account the great increase in the unstructured labour market that is knowledge-based, highly skilled and specialized (Beck, 2003). Beck also adds that:
The international labor organization reported that companies in industrialized countries are likely to depend increasingly on recruiting their skilled manpower needs from the external labor market. (Beck, 2003)
Based on the human capital theory, Hersch (1991) suggested that education is a highly visible indicator of work quality. As such, employers screening workers for jobs which do not require specific training may find the cheapest way to screen prospective employees is the worker’s education. Higher worker education implies higher productivity (Becker, 1964). A prior study has also suggested that a contingent workforce is needed to reduce overheads (Gimble, 1991). However, in practice, these theories may be not applicable for the case of Vietnam because not all jobs require a high worker-education level. Typically employers use methods which will attract the required number of employees during the main business periods, rather than being primarily concerned with the quality of the workforce.
Firm age
The age of the firm may be a potential moderator of a firm’s financial value generated by managers. In particular, in a young organization the opportunities for executive-level involvement in activities across the organizational hierarchy are numerous and varied. More specifically, a founder CEO will characteristically be highly involved in most, if not all, significant aspects of their organizations’ functioning (Jayaraman et al., 2000).
The relationship in economic theory between firm size and growth has traditionally involved the two notions of the optimum size of the firm and industrial equilibrium. The former has been rigorously restated recently with respect to specific propositions made by organizational theory. In accordance with traditional theory, the optimum size is mentioned in relation to the assumption that all firms deal with the same U-shaped long-term average cost curve. In fact, large firms would be expected to be at or near their optimum size and would therefore grow very little and might even shrink if they exceed optimum size. Small firms on the other hand would be far below the optimum size and would need to grow at a faster rate to achieve this size (Singh and Whittington, 1975). Previous research has claimed that firm size plays an important role in explaining firm performance. Specifically, large firms enjoy advantages such as lower costs and higher returns due to greater access to the capital market (Hall and Weiss, 1967) and economies of scale (Montgomery, 1985). It should be noted that they used total asset value as an indicator of size in their measurement of performance, while we will use the number of employees to measure the effects of size on firm performance (Jayaraman, 2000).
Ownership
Ownership structure may influence firm performance. For instance, with substantial ownership of cash flow rights, family ownership has the incentives and power to take actions that benefit themselves at the expense of firm performance. In contrast, diverse shareholders are presumed to evaluate investments using market value rules that maximize the value of the firm’s residual cash flows. Large numbers of concentrated shareholders, however, may derive greater benefits from pursuing objectives such as firm growth, technological innovation, or firm survival rather than enhancing shareholder value (Anderson and Reeb, 2003), or may have substantial economic incentives to diminish agency conflicts and maximize firm value (Demsetz and Lehn, 1985). However, one of the greatest costs that a group of concentrated large shareholders can impose arises if they remain active in management when they are no longer competent or qualified to do so (Shleifer and Vishny, 1997).
Bribery
public official (World Bank, 2000). As mentioned above, in relation to corruption, a distinction can be made between administrative or bureaucratic corruption, which refers to paying bribes for services concerning the implementation of regulations, and state capture, where firms try to influence the formulation of laws, regulations, decrees and other government policies to their own advantage through illicit or non-transparent means (Fries et al., 2003). Only the first category is relevant to the present context as private firms must acquire the consent of government officials to carry out their investment plans (Ninh et al., 2007). We do not examine the illegal misappropriation of public property by public officials, although sometimes this is also seen as part of the corruption phenomenon (Gaviria, 2002).
Industry context and firm performance Industry sector
Industry effects are vital in the analysis of firm performance. Firm performance is a function of operating in profitable industries (Schmalensee, 1985). Some of the early research in industrial organization concluded that differences in profitability across firms could largely be explained by industry membership and that industry performance levels could be explained by barriers to entry and other structural characteristics (Scherer, 1980). Firms in new, expanding industries are expected to perform better than those operating in old, declining industries. Business firms in a particular industry may gain benefits that are comparatively higher than normal profits because of certain attributes of the economy of the country (Sarkaria, 1997), or some structural variables may allow firms in particular industries to be in a better position to implement their strategies successfully and profitably (Pant, 1991).
Level of competition
firms. Such markets are large and demand grows rapidly. Products are commercially viable and customers are aware of the product advantages. Hence, there is ample room for a young firm’s entry. Additionally, growth markets are characterized by turbulence because the competitive structure is changing. Established firms are often locked into old technology and related market approaches due to their previous investments. In other words, mature markets are large markets with stable or slowly growing demand. In mature markets there are less changes in competition, products and technology than in growth markets (Anderson and Zeithaml, 1984). Customers are familiar with the products and services within the industry, and market share is stable (Hambrick, MacMillan and Day, 1982). Consequently, mature markets provide limited opportunities for new firms because these have few, if any, advantages over their established competitors. In short, high levels of competition, especially from large and well-established competitors, limit the growth opportunities of young firms.
2.3 Hypotheses
This study investigates four demographic variables of CEOs, namely tenure, formal education, informal education and CEO risk aversion. Additionally, we examine the impact of three organizational characteristics on firm performance, namely employee education, firm age and bribery. We control for focal firm size, level of competition, industry and type of ownership.
2.3.1 CEO characteristics
investment required to keep the firm evolving over time. Reuber and Fisher (1999) therefore emphasized that researchers should understand the diminishing returns of experience, rather than accept the linearity of the relationship between experience and performance. Hambrick and Fukutomi (1991) posited an inverse U-shaped relationship between tenure and firm performance among CEOs whose tenures span a long period. Chandler (1996) found a curvilinear relationship between the depth and breadth of experience6 and a similarity between skills and ability, as well as sales growth. This brings us to our first hypothesis:
Hypothesis 1. There is a curvilinear relationship between CEO tenure and firm performance.
Formal education. The educational background of managers is an important variable to be studied. The level of formal education measures an individual’s knowledge and competence base (Hambrick and Mason, 1984; Hitt and Tyler, 1991). CEOs with higher levels of formal education are expected to generate a wider range of creative solutions when faced with complex problems. The level of formal education has been positively linked to firm performance (Norburn and Birley, 1988), the degree of firm innovation (Bantel and Jackson, 1989), and change in corporate strategy (Wiersema and Bantel, 1992). It has been suggested that the level of formal education attained by managers is an important factor that affects the strategy development process. Karami et al. (2006) showed that well-educated CEOs are inclined to develop formal strategic plans. The strategic awareness of CEOs plays a significant part in the formulation of business strategies. We expect that companies with CEOs who have a high formal education background will be more inclined to achieve higher firm performance than companies with managers who have not benefited from formal education. We therefore hypothesize as follows:
Hypothesis 2. There will be a positive relationship between CEO formal education and firm performance.
6
Informal education. Investments in human capital, such as attending training courses, leads to higher wages − presumably because of higher productivity (Becker, 1964). Increasing educational attainment is positively related to receptivity towards innovation, regardless of the form of education (Barker and Mueller, 2002). There are few empirical studies on the influence of the informal education of managers, and the evidence regarding their effects on firm performance is sparse. As such, the effects of informal education on firm performance have barely been examined. We may reasonably expect contingency effects: high levels of informal education are more conducive to organizational performance in some environments, and in pursuit of some strategies, than in others. We therefore expect the following:
Hypothesis 3. There will a positive relationship between CEO informal education and firm performance.
environments, risky decisions or changes of strategies may be necessary for success. This leads to the following hypothesis:
Hypothesis 4. There will be a negatively significant relationship between risk-averse CEOs and firm performance.
2.3.2 Firm characteristics
Firm age. Prior research has found company age and firm performance to be negatively related (Begley and Boyd, 1985; Dunne and Hughes, 1994). Questioning Gibrat’s law of proportionate growth,7 some researchers have found a negative size-growth relationship (Wagner, 1992; Hart and Oulton, 1996). The inverse relationship between age and performance may also be due to all the weaknesses that may arise with ageing (Shergill and Raman, 2005). We suggest that older firms may have lower performance levels than younger ones because of the continued use of outdated management and marketing practices and/or obsolete technology and their in-built resistance to new approaches. Firm age may have the effect of limiting the ‘required’ involvement of founder CEOs, and one might consequently expect a negative relationship between the strength of the effects of founder management and firm age (Jayaraman et al., 2000). We therefore hypothesize that:
Hypothesis 5. There will be a negatively significant relationship between firm age and firm performance.
Employee education: The human capital of workers is expected to enhance firm performance due to their skills, experiences, self-employment and education possibly contributing to increased productivity and yielding knowledge about profitable strategies (Becker, 1964), and a real difference in productivity connected with one form of human investment, that is, education
7
(Theodore, 1961). For these reasons, we predict there will be a positive association between employees’ education level and firm performance. We therefore hypothesize that:
Hypothesis 6. There will be a positively significant relationship between employee education and firm performance.
Bribery. Corruption (‘greasing the wheels’) causes an increase in business costs for enterprises due to the price of illicit payments made by them (firms or agents) to public officials (the principal) (Maitland, 2001), the management costs of negotiating with officials and the risk of violation of agreements or detection of these agreements (Ninh et al., 2007). Maitland (2001) claimed that corruption is viewed as a key constraint, holding back economic growth, particularly through its negative impact on investment incentives for domestic private enterprises. The bribe represents the transfer of part of the income stream (rent, for example) from the asset or assets that the individual seeks to exploit. Nevertheless, bribery can also benefit private firms seeking to borrow adequate funds − equivalent to their mortgage assets from local banks − because if bankers accept bribes, they may well ignore an accurate appraisal of a customer’s debt-repaying capability or the real value of the mortgage asset. Conversely, without bribes they may limit borrowing to 50 percent of the total mortgage asset (Ninh, et al., 2007). Ninh et al. (2007) also found an inverted U-shape curve between bribes and firm investment. In accordance with that result, we expect that there will be an inverted U-shaped association between firm performance and bribery, meaning that firms’ willingness to pay bribes in an effort to expand investment has a determined limitation because if the amount of money for a bribe exceeds profit value (certain point) for any specific investment opportunity, or incurs a burden cost, firm performance will decline. This leads to the following hypothesis:
2.3.3 Control variables
We included four control variables concerning firm performance in our analysis. These variables are recognized as having an influence on firm performance. They are firm size, level of competition, type of ownership and industry sector. We include firm size (number of employees) as a variable in our model to control for internal factors which may have an influence on financial outcome. Secondly, we assess the level of competition as an external factor directly impacting on firm performance. A similar argument applies to the industry sector, which seems to have a different profit level operating in attracting firms to enter. Our final control variable is the type of ownership, which differs across firms. We consider private, joint-stock companies, limited-liability firms, collectives (cooperatives), and household or family-owned businesses.
2.4 The theoretical model
We will use the OLS regression method to estimate values in a Log-linear model in order to determine the interaction of independent variables with respect to dependent variable.
ε β α + + = iXi LnY Where Y: Income (LnSales) X1: CEO tenure
X2: CEO formal education (dummy)
X3: CEO informal education
X4: CEO risk averse (dummy)
X5: Employees’ education
X6: Firm age
X7: Money to bribe
X8: Firm size (number of employees)
X9: Private (dummy)
X10: Service (dummy)
3. RESEARCH METHODS
3.1 Data collection and sample
This research is a part of our ongoing research project, and was conducted from the beginning of 2003 until March 2005. However, for the cross-sectional data analysis which was chosen for this thesis, we decided to only use the data collected for 2004, due to the fact that this year provided the most complete information. Secondary data can be easily collected for each province using local administrative offices such as those concerned with statistics, investment and tax, but this data is often aggregated and thus is not applicable at the firm level. For this reason, we used a questionnaire to collect the information, which included questions concerning issues such as the firm’s age, the CEO’s education level and sales figures. Specifically, a questionnaire containing 35 questions was developed which was designed with the help of contributions from several professional advisers, and derived from many different sources. The data we collected includes much diverse information about CEO characteristics and also firm organizational traits and firm investment. The research population was firms in six provinces of the MRD in Vietnam. The reason we concentrated on this region was because, as mentioned above, it shows a significant increase in the number of private firms in recent years, the performance of which is differently reflected in profits. Additionally, the key role of private firms in the MRD contributes greatly to the GDP of the entire country.
Our research proceeded in three stages. In the preparatory phase of the fieldwork, we first implemented several pilot surveys in two provinces, Can Tho and Kien Giang. This resulted in a number of modifications to the questionnaire. In the second stage, a team of interviewers was trained, consisting of teachers and students from the School of Economics and Business Administration (SEBA), Can Tho University, Vietnam. The students selected were required to have some experience in conducting surveys. Of course, we also trained the selected students prior to conducting the questionnaire. We made them aware of the importance of the data they would be collecting, with the intention of motivating the students to take responsibility for the data collection as a means of improving data quality. In addition, the interviews were conducted in the local language, which interviewees respond to more easily, making their answers more precise.
In the third stage, we conducted direct interviews with 606 managers of private firms. We did not select a sample prior to the interviews, rather, the sample was selected randomly on the basis of the owners being willing to cooperate. If they agreed, we started to interview them, whereas if they refused we apologized and proceeded to the next firm. Based on experiences drawn from previous research, we supposed that it might be risky to design for a fixed quantity of samples at first that would be based on firms we already knew of from publicly available information, websites and local offices. This was because this method might not supply enough samples, due to the possible absence of the owners,8 or the CEO’s refusal to respond (interviewees refused for reasons such as not wishing to disclose information, being too busy, or feeling uncomfortable when being asked about their business). Instead of adopting this approach we only provided a questionnaire to the firms encountered whose owners agreed to an interview. In fact, the questionnaire was conducted only if the owner or CEO was available to answer personally in order to obtain complete and correct information. If the prospective interviewees were absent, we left the questionnaire and returned having made an appointment. Some benefits resulted from this method, as we obtained more than the necessary samples, and also saved time and reduced our substantial transportation costs. During the interview, the main topics, such as tenure, firm age, education, investment, loans and industry context were
8
discussed. Some extra questions were added to invigorate the interview and to enable the respondents to tell their own story to some extent.
We obtained 606 useable responses from the total number of firms contacted. Although the non-response and reasons for not participating in the survey, as noted earlier, are not recorded, we believe the quality of the questionnaire and the way it was conducted should instil strong confidence in the quality of the data set. Finally, as mentioned above, as a result of the need for cost efficiency we restricted the survey to six of the thirteen provinces of the MRD.
3.2 Measures
Table 4 gives an overview of the items we used for a particular construct. Without discussing these at length, we ensured that the content of each item – the wording and phrasing of the particular question in the survey – directly related to the construct it was designed to measure (‘face validity’).
[Insert TABLE 4 about here]
First, we measured firm performance by the natural logarithm of the firm’s sales (in Vietnamese dong). We used the natural log to obtain a normal distribution of firm sales (see Figure 1, appendix).
Second, we used four items to measure the effects of CEO demographic variables on firm performance.
CEO Tenure (CEOTENU). CEO tenure was measured by the total number of years the CEO had worked for both the focal firm (Hill and Phan, 1991) and tenure at any other firms (Hambrick and Fukutomi, 1991). It should be noted that the longer a CEO had held a management position, the more managerial experience was gained. Moreover, Ellen Vanbilsen et al. (2002) suggested that there are three dimensions of management experience, entrepreneurial, industrial and managerial. In fact, managerial experience is the total experience in management regardless of the industry (Lee and Tsang, 2001).
CEO informal education (CEOINFOR). CEO informal education was measured by the number of times CEO participated in management training courses. It covers informal training programmes or “informal training” as short training courses to update knowledge in fields such as management, marketing, and accounting.
CEO risk averse (CEO_RISK). We used a dummy variable that equals 1 if a CEO is risk averse, 0 otherwise. It means that a CEO’s ability to take risks depends on his preferences. In fact, to assess firms’ attitudes towards risk we asked the CEO to choose between two possibilities: (i) investing VND 1 million to earn an assured profit of VND 100,000 and (ii) investing VND 1 million to earn a profit of VND 200,000 with a 50 percent probability of success. The statistical result shows that 81.7 percent of the CEOs in the sample are risk averse.
Third, three items were used to measure the effects of firm characteristics on firm performance.
Firm age (FIRMAGE). The length of time the company has been established. The age of the company was calculated by subtracting the year the firm was founded from the current year (Ellen Vanbilsen et al., 2002).
Employee education (EM_EDU). This is not similar to the CEO formal education measurement. The education level of the employees was measured by a continuous variable (formal education). It indicates whether employees have a degree or have attended a formal educational institution.
Bribery (MBRIBERY). Bribery refers to the amount of money firms use to bribe local officials in order to gain ‘benefits’ in their business dealings. These include bribes that businesses pay to government officials in order to speed up bureaucratic processes and reduce the time spent negotiating with officials (Ninh et al., 2007).
Finally, we used four control variables to measure extra influences on firm performance. Level of competition (LEVECOMP). Level of competition was measured by a dummy variable that equals 0 if the CEO suggests that the company operates in an industry with a low-medium competition level, 1 otherwise.
Industry sector (SERVICES). Industry sector was also measured by a dummy variable that equals 1 if the firm is a service business, 0 otherwise (exploiting-processing-manufacturing and trade sector).
Firm size (EMPLOYEES). Firm size was measured by the number of employees who (frequently) worked for the company. This is also considered by other studies to be an indicator of firm size (Dejong and Woolthuis, 2007; Tihanyi et al., 2000; Datta and Guthrie, 1994). However, other studies have measured firm size by total assets. For instance, Vanbilsen et al. (2002) used the average of total assets to reflect firm size. Other researchers have employed the log of total assets as the measure of size (e.g. Siddharthan and Dasgupta, 1983; McKnight and Tomkins, 1999). However, in this study we only use the number of employees to measure firm size.
3.3 Analysis
The analyses were carried out on the set of firms whose sales are available for only one year (2004) with the sample size of 606 observations. The impact of CEO demography and firm characteristics on firm performance (lnsale) were investigated using OLS models on cross-sectional data. Two sets of OLS regression models were performed in order to test the hypotheses. The sample includes missing observations for particular items. In one sample we deleted all missing values. In another sample we replaced the missing value by an estimated mean value calculated using the expectation-maximization method (EM).9 By doing so, we had a full observation. We called the former, with the missing value, the incomplete model (Model 1) and the latter the robust model (Model 2).
In Model 1, given a valid sample of 401 observations, we first performed a regression on firm outcomes (lnsale), along with the control variables for Model 1, to test for minor effects of a set of contextual factors (employees, private, services, and competition) on firm performance. Second, we conducted a similar regression adding only CEO demography variables together with all control variables in order to test the impact of CEO personal characteristics (formal and informal education, CEO tenure, CEO risk averse) on firm outcomes. In particular, for this step, we also examined whether there would be an inverted U-shaped
9
relationship between CEO tenure and firm performance by adding the squared term of CEO tenure into the model. Third, we only investigated firm characteristics along with a set of control variables. Similar to step 2, we also checked for an inverted U-shaped relationship between firm age and firm performance. At step 4 we included all variables (along with control variables) into the model in order to test the total effects of both CEO demography and firm traits on firm outcomes. Finally, we removed two squared terms which may be not good predictors and statistically insignificant in the OLS model. Consequently, we used this adjusted OLS model for testing whether the OLS model is the best linear unbiased estimator (BLUE).
In Model 2, we proceeded with the OLS regression model for all 606 valid observations. Similar to the steps in Model 1, we gradually added CEO traits and firm characteristics along with control variables. Eventually we tested an adjusted model. More importantly, to test the effects of a set of all variables added into the model, we also reported R2 increments in the OLS regression model. By so doing, we could avoid the problematic biases of specific estimated coefficients.
4. EMPIRICAL RESULTS
4.1 Sample characteristics
0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00%
Private Limited Stock Collective Households
0 --> 500 500 ---> 10,000 > 10,000
Figure 2. Sales and ownership
In group 2 − sales ranging from VND 500 to 10,000 million − private firms (sole proprietorship) again prevail with the highest percentage (55.8%); households or family-owned businesses (23.3%); limited-liability companies (16.3%); stock or joint-stock companies (3.1%); collectives or cooperatives (1.6%). However, if we compare each ownership type across the three different group sales categories, ownership types other than households predominantly occur in this group, including private (46%), limited (60%), stock (57.1%), and collective (66.7%). In group 3 − sales more than VND 10,000 million − although few private firms (7%) have total sales larger than VND 10,000 million, they still account for the highest percentage of firms with sales over VND 10,000 million (56.4%).
49.41% 44.55% 6.04% Year 1 --> 5 Year 6 ---> 20 Year > 20
Figure 3. CEO tenure in Vietnamese private enterprises
Figure 4 reveals that the vast majority of CEO education occurs at high school (48.1%) and university undergraduate degrees or higher (23.8%). In addition, the sample distribution shows that some CEOs have only secondary level (20.5%), or even primary level education (7.6%). Not surprisingly, CEOs with low levels of education, unless managerial training courses were very popular within their firms, found it very difficult to participate in such courses and update their knowledge, which in turn limited their ability to implement innovation strategies and capture necessary information.
7.60% 20.50% 48.10% 23.80% Primary Secondary High Shool University or above
Figure 5 shows the number of enterprises categorized by the types of ownerships, in which private firms account for the highest percentage ratio (51.7%), households (33.5%), limited companies (11.6%), stock (2.3%) and collective (1.0%). Notably, one of the special characteristics of ownership type for households is family-owned business. The CEO is often an owner and founder. Interestingly, this type of business is similar to private firms but its business scale is smaller. Typically, it may employ staff, but not very many, and usually family members, other relatives, or itinerant labours.
51.7% 11.6% 2.3% 1% 33.5% Private Limited Stock Collective Households
Figure 5. Ownership types
The sample distribution of firm age spread and standard deviation is quite high (7.541). The age of most firms is in the range of 1 to 5 years, accounting for 336 enterprises (55.4%); 236 enterprises were 6 to 20 years old (38.9%); while 34 enterprises were more than 20 years old (5.6%) (see Figure 6).
336 236 34 Age 1 ---> 5 Age 6 ----> 20 Age > 20
Figure 7 reports that most private firms concentrate mainly on trade (304 enterprises, 50.4%); while exploiting-processing-manufacturing accounts for 34.3%, services account for is 15.3% of the total sample. In particular, in Kien Giang and Can Tho provinces the trade sector predominates as opposed to other sectors, 34.2% and 26.3%, respectively.
[Insert FIGURE 7 and TABLE 5 about here]
4.2 Description
Table 6 reports the means, standard deviations, minimum and maximum values, skewness and kurtorsis. In 2004 the average total sale (LNSALE) of the sample is 6.39. A standard deviation of 1.86 implies a small variation in sales across the firms after taking a natural logarithm. The maximum and minimum values are 13.92 and 0.69, respectively. Given skewness (0.452) and kurtosis (1.039), the distribution of sales is nearly normal. The average tenure of CEOs (CEOTENU) is 8.35 years. A standard deviation of 7.07 reflects a large variation of tenure across the firms. The longest tenure for any CEO is 44 years. The average formal education level of CEOs (CEO_EDU) is high school while some CEOs have a very low level of education. Surprisingly, several CEOs had never participated in management training courses (CEOINFOR) to enhance their management skills and knowledge, while others had attended courses up to 16 times. The average the number of times a CEO participated in a course is at least once (s.d 2.605). In addition, a skewness of 3.955 and kurtosis of 17.391 imply that the distribution of informal education has a long right tail and a high degree of concentration. In practice, when conducting the survey, we were quite amazed because some CEOs proposed that some training courses were both inefficient and a waste of their time and money. CEOs (CEO_RISK) who are risk averse account for 81.7%, while those who are risk-takers account for 10.4% of 558 observations (See Table 16, appendix).
moderately high degree of concentration. The number of employees (EMPLOYEES) indicates an average value of 25.6 (s.d 152.76), the maximum and minimum values equal 2500 and 1 respectively. The distribution of EMPLOYEES depicts a long right tail and high degree of concentration, given skewness (14.782) and kurtosis (234.099). According to the data set, the average amount that the firms have to pay in bribes is VND 1,418.30 thousand per firm with a standard deviation of VND 9,162.67, that is, 6.5 times the average, implying that the amount of bribes paid also varies across firms. With a skewness of 13.138 and a kurtosis of 204.637 the distribution of MBRIBERY has a long right tail and a high degree of concentration.
TABLE 6. DESCRIPTIVE STATISTICS
N Minimum Maximum Mean
Std.
Deviation Skewness Kurtosis
LNSALE 584 .69 13.92 6.389 1.863 .452 1.039 CEOTENU 514 0 44 8.35 7.076 1.435 2.292 CEO_EDU 606 0 1 .61 .489 -.433 -1.818 CEOINFOR 552 0 16 1.03 2.605 3.955 17.391 CEO_RISK 606 0 1 .816 .387 -1.642 .699 FIRMAGE 606 0 62 7.59 7.541 2.419 8.989 EM_EDU 585 1 3 1.41 .621 1.254 .461 MBRIBERY 506 0 162,000 1,418.30 9,162.675 13.138 204.637 EMPLOYEES 591 1 2500 25.63 152.762 14.782 234.099 LEVECOMP 606 0 1 .67 .471 -.717 -1.491 SERVICES 606 0 1 .15 .359 1.945 1.791 PRIVATE 606 0 1 .52 .500 -.066 -2.002
Source: Own survey, 2005
firms. In fact, whenever a new firm enters a market its sales often increase very rapidly in the early stages because of the predatory price policy used in order to gain market share, meaning that older firms are unable to maintain their share. Employee education is positively associated with firm performance while the number of employees (firm size) has quite a high positive correlation with firm sales. In practice, most firms (exploiting-processing-manufacturing) require a large number of employees (itinerant or local) who only work for the firms temporarily. Especially during the peak time of a seasonal business, the firms found it difficult to hire enough workers due to local competition for labour, or from jobs offered in other provinces. Bribery also has a positive correlation with firm sales, implying that bribery, in some cases, does help improve firm performance. We will provide further explanation in the following sections.
TABLE 7. CORRELATION MATRIX (N=401)
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) 1.LNSALE 1 2.CEOTENU ...015... 1 3.CEO_EDU .214** -.189** 1 4.CEOINFOR .206** -.007... .055... 1 5.CEO_RISK -.069... .094... -.024... .059... 1 6.FIRMAGE -.111*.. ...636** -.272** .056... .100* 1 7.EM_EDU .-.153**. .-.122*. .266** ..172** -.107* -.138** 1 8.MBRIBERY ..147** -.007.. .122*. .061... .046 -.078... .249** 1 9.EMPLOYEES ..298** .056.. .115*. .178** .050 -.030… .061… .169** 1 10.PRIVATE ..104*. .-.105*. .057.. .040... -.112* -.091... -.084…. -.014... -.064.… 1 11.SERVICES .-.092... .-.080.. .043.. .069... .057 -.142** .146** .090.. .173** .009.. 1 12.LEVECOMP ...126*. ..023.. .013.. -.012... -.088 -.057.. .001… .031.. -.017… .115* .006 1 ** Correlation is significant at the 0.01 level (2-tailed).
* Correlation is significant at the 0.05 level (2-tailed).
4.3 Regression results
Table 8 summarizes the results of the regression analyses testing the impact of CEO demographic characteristics and firm characteristics on firm outcomes. By using the procedures described in the analysis section, we obtained all the statistically significant regression models (Model 1: F = 21.16, p < 0.01; Model 2: F = 14.24, p < 0.01; Model 3: F = 11.98, p < 0.01; Model 4: F = 8.67, p < 0.01; Model 5: F = 9.99, p < 0.01). Model 1 includes only control variables. Model 2 estimates the impact of CEO demographic characteristics together with the control variables on firm performance. Model 3 tests the influence of firm characteristics along with control variables on firm outcomes. Model 4 consists of all variables including squared terms. Model 5 reports a selected model after removing the squared term. This hierarchical approach allows us to assess the unique contribution of both CEO demography and firm characteristics on firm outcomes. In addition, in order to obtain a BLUE, we tested the assumptions of the regression models. The results report that there is no heteroskedasticity (the assumption var =
σ
2is not violated), multicollinearity, or serial autocorrelation in the model even if residuals of the estimated model have a non-normal distribution (p-value of Jarque-Bera test < 0.01) (see Research methodology paper). Therefore, we conclude that the least squares estimator is BLUE and the confidence interval for the estimated coefficients is reliable.Results from Model 4 report that (1) firm performance will be improved by an increase in firm size (the number of employees) (p < 0.01); (2) the service sector may have lower sales than other sectors (trade; exploiting-processing-manufacturing) (p < 0.01); (3) private firms (sole proprietorship) may have greater sales than others (limited, stock, collective and household) (p < 0.05); (4) it seems to be the case that when a firm is located in a highly competitive environment firm sales increased more than in a low competitive environment (p < 0.05).
performance, all other things being equal. Specifically, CEOs who are highly educated tend to invest more because they can access and exploit investment opportunities more easily than those who do not. In addition, they may have more opportunity to access loans due to a better understanding of bank regulations and, simultaneously, may establish a wider network of relationships (Ninh et al., 2007). The results provide no evidence for Hypothesis 4, implying a negative association between CEO risk-averse behaviour and firm performance (β = -0.248, n.s). Thus, CEO risk aversion is not significantly related to firm performance.
Hypothesis 5, that firm age is expected to have a negative relationship with firm performance, is also not significantly statistically supported (β = -0.026, n.s), although it is supported in Model 5. Hypothesis 6, indicating a positive relationship between employees’ education and firm outcome is statistically supported (β = 0.265, p < 0.10). Our results actually suggest that employees with a higher education find it easy to find specialized work due to their improved cognitive capability, and as a result are indirectly better for firm performance. Hypothesis 7, an inverse U-shaped relationship between the amount paid in bribes and firm sales, is supported. The results provide strong evidence for this hypothesis, that is, MBRIBERY has a positively significant relationship with firm sales (β = 0.000, p < 0.01), while SQMBRIBERY (squared term) has a negatively significant relationship with firm outcome (β = -1.33E-09, p < 0.05). As expected, when the average amount spent on bribes exceeds a certain limit, the desire to invest in the firm will decline, and as a result firm performance decreases.
As for the control variables, our results suggest that firm size is significantly related to firm performance (β = 0.008, p < 0.01). The same applies to the level of competition. The higher the level of competition in industry, the more firm sales can rise (β = 0.430, p < 0.05). We also found evidence showing that privately owned firms have greater sales than other ownership types (β = 0.348, p < 0.05). Service industry sales are likely to be smaller than other industry sales figures. This relationship is confirmed (β = -0.988, p < 0.01).
TABLE 8.REGRESSION RESULTS
Dependent Variable: LNSALE Included observations: 401 Estimated coefficient (t-statistic)
Variable Model 1 (Control variables) Model 2 (CEO demography) Model 3 (Firm characteristics) Model 4 ( Full model) Model 5 (Adjusted Full Model) C 5.931*** (40.23) 5.460*** (17.94) 5.391*** (18.25) 5.134*** (12.37) 5.346*** (14.16) Control variables EMPLOYEES (FIRM SIZE) 0.004***
4.4 Robustness analysis
According to the data set, given missing values from the 606 observations, leaving only 401 complete observations, we realize that bias is possible. As a result, for robustness we used the EM method which adds (imputes) missing values by estimating mean values (see Research methodology paper). We also used the same procedures of regression analysis as was undertaken with the analysis of the incomplete model. Table 9 reports the results of the regression analyses.
[Insert TABLE 9 about here]
We have considered the robustness of our results in the face of a number of other estimation problems, including heteroskedasticity, serial correlation, and collinearity with the aim of testing whether the assumptions of the OLS model were violated or not. Results reveal that there is no heteroskedasticity, multicollinearity or serial autocorrelation in the model. Therefore, the least squares estimator is BLUE (see Research Methodology paper).
5. CONCLUSIONS
5.1 Conclusions
In this section, we will discuss and compare prior research of relevance to our study in order to gain insights into the effects of CEO demography characteristics and organizational traits on firm performance. The most striking hypotheses and outcomes are summarized (see Table 10).
TABLE 10. SUMMARY OF EMPIRICAL FINDINGS
Hypotheses Prediction
Result N=401
Result N=606
1. CEO tenure – performance ∩ ∩(*) ∩
2. Formal education – performance + + +
3. Informal education - performance + + +
4. CEO risk averse – performance - - (*) - (*)
5. Firm age - performance - - -
6. Employees’ education - performance + + + (*)
7. Bribery – performance ∩ ∩ ∩
*ns: not significant at p<0.05 or p<0.10
Source: Own survey, 2005
First, based on the mixed results, we argue that the a long-tenured CEO results in poorer performance as compared to the a short-tenured CEO. The estimate is consistent with some prior studies as well as the theoretical basis. For example, Finkelstein and Hambrick (1990) reported that long-tenured managers are risk averse as they have far more to lose than to gain from changing the current equilibrium. This implies that short-tenured CEOs have more attributes likely to improve firm performance than long-tenured CEOs. Ken et al. (1994) found a positive relationship between team tenure and sales growth or ROI, based on the rationale that greater tenure facilitates both coordination and control, but this relationship does not show a statistical significance. There may be other reasons why long-tenured CEOs have a negative influence on firm performance. For instance, after a long tenure, CEOs seem to become overconfident and complacent, resulting in rigidity and resistance to reorientation and external pressure for change (Meyer, 1975). This does not mean that long-tenured CEOs will necessarily resist change but their changes may be consistent with past traditions and the direction of evolution may entail a divergence or mismatch of the organization and its environment. As previous studies suggest, with each year in the position, the CEOs within our survey developed a distinctive, standard way of processing information, including that arising internally within the industry, as well as that arising from clients or suppliers, with the sources of his or her information becoming increasingly narrow and restricted. Limited access to public information may be an important factor that indirectly affects a CEO’s ability to obtain information. This leads to CEOs tending to rely increasingly on their past experience or habits, rather than utilizing new information.
