Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

The Investment Opportunity Set and the Voluntary Use of Outside by etssetcf


The Investment Opportunity Set and the Voluntary Use of Outside

More Info

 EBMS Working Paper EBMS/2000/8
       (ISSN: 1470-2398)

  The Investment Opportunity Set
     and the Voluntary Use of
         Outside Directors:
       New Zealand Evidence

M. Hossain, S. F. Cahan, M. B. Adams

   European Business Management School
              Singleton Park
             Swansea SA2 8PP

              Tel: 01792 295601
      (International: +44 1792 295601)
              Fax: 01792 295626
      (International: +44 1792 295626)
    The Investment Opportunity Set and the Voluntary Use of Outside
                   Directors: New Zealand Evidence

M. Hossain, S. F. Cahan, M. B. Adams*

Abstract — This study examines whether the composition of boards of directors differs between high and
low growth firms. Based on prior research, we hypothesise that firms with greater investment
opportunities require more monitoring because managers in these firms have more discretion both in
selecting investments and allocating resources between investments. Because outside directors can be
more effective monitors than inside directors, we predict that outsiders will make-up a larger proportion
of the board in high growth firms than in low growth firms. Using a cross-sectional sample of 77 New
Zealand firms, our results suggest that the percentage of outside directors is related to growth for two
of the four measures of investment opportunities which we employ. As expected, the percentage of
outside directors is also related to a composite measure of investment opportunities.

   Mahmud Hossain is Assistant Professor at Nanyang Technological University. Steven F. Cahan is Professor at
Massey University. Michael B. Adams is Professor at the University of Wales - Swansea. The authors appreciate
the comments of workshop participants at University of Exeter, Macquire University and Nanyang Technological
University. Financial support received from the Massey University Research Fund is also very much appreciated.
1. Introduction

The corporate governance role played by outside directors has been a topic of considerable controversy
in both the academic and professional literature (e.g., Fama and Jensen, 1983; Bhagat and Black, 1997;
Shleifer and Vishny, 1997; Klein, 1998). Fama and Jensen (1983), among others, argue that independent
or outside directors can efficiently and effectively control management in agency settings where
ownership and management are separate. For example, Klein (1998, p. 301) reports that "firms
increasingly are replacing their inside directors with outsiders ... primarily [due] to external pressures
from groups (for example, shareholder activists and the financial press) who actively advocate the
perceived benefit of ... firms having boards totally independent from management." Therefore, outside
directors are valued because their interests are more closely aligned with the owners when compared to
inside directors who hold positions in the firm and who may have incentive to pursue activities that do
not increase firm value.1

Numerous studies (e.g., Byrd and Hickman, 1992; Brickley et al., 1994; Agrawal and Knoeber, 1996)
examine the determinants of board composition and/or the effectiveness of outside directors.2 However,
none of these studies focus explicitly on the link between a firm's investment opportunities and the
composition of its board. This is surprising as board structure is likely to be endogenously determined
and is likely to be related to the firm's growth options. For example, prior research (e.g., Smith and
Watts, 1992; Gaver and Gaver, 1993; Skinner, 1993) find that contracting mechanisms related to capital
structure, dividend policy, compensation and accounting policies are related to the firm's investment
opportunity set (IOS). In order words, because there are cross-sectional differences in investment options
and different investment options imply variations in the levels and types of agency costs, firms will
choose several contracting mechanisms to efficiently address these costs.

We argue that board structure is another contracting mechanism that can be used to reduce agency costs.
Accordingly, we expect to find cross-sectional variation in the proportion of outside directors on the
board, and this variation should be a function of the firm's IOS. Using a final sample of 77 New Zealand-
based firms, we find that the proportion of outside directors is positively related to two of three common
measures of IOS. In addition, we find that board composition is inversely related to control variables for
low inside ownership and firm size and positively related to control variables for leverage and the number
of board meetings per year. In contrast, a control variable for Chief Executive Officer (CEO) tenure does

                  Others question the ability of outside directors to objectively monitor corporate
decisions. For example, Patton and Baker (1987) point out that outside directors are usually nominated
and appointed by top management, and often hold only a trivial fraction of the firm's shares.

                  See Bhagat and Black (1997) and Shleifer and Vishny (1997) for reviews of this
not appear to be an important determinant of board composition for our sample. Overall, our results add
to the literature by showing that corporate governance is another managerial policy issue that is affected
by IOS.

We divide the rest of the study into five more sections. Section 2 develops our hypothesis. Section 3
provides background information on the institutional environment in New Zealand. Section 4 describes
the research design, including the sources of data, the statistical model used and the measurement of the
variables. Section 5 presents the results, while section 6 summarises and concludes the paper.

2. Hypothesis

Myers (1977) argues that firm value consists of two main elements. These are: 1/ real assets (called
assets-in-place) which are valued independently of managers' future discretionary investment, and 2/ real
options (assets yet to be acquired) where the value of real options depends on future discretionary
investments. Examples of the latter include expansion projects, new products, business acquisitions, and
marketing programmes (Gaver and Gaver, 1993).

Alchian and Woodward (1988) define IOS differently than Myers (1977). They view IOS as being related
to a resource's "plasticity" and monitoring costs. They defined investments as plastic if they have "a wide
range of discretionary, legitimate decisions within which the user may choose or that an observer can less
readily monitor the choice" (Alchian and Woodward, 1988, p. 69). In comparing drug and steel
manufacturing firms in the United States (US), Alchian and Woodward (1988) indicate that the former
had wider initial options to control decisions about resources than the latter. The activities of drug
companies may also be difficult to monitor and control because of the specialised nature of their

When managers have special information or specific knowledge about investment options, it can be
efficient to let them choose which options to pursue (e.g., Jensen and Meckling, 1995). However, in such
cases, agency costs also increase because managers may not always have incentives to maximise firm
value (e.g., see Brickley et al., 1997). For example, managers may have incentive to empire-build by
making diversified acquisitions that reduce firm value. Gaver and Gaver (1993, p. 129) write:
     "[A]s the proportion of firm value represented by growth opportunities (as opposed to assets in
     place) increases, the observability of managerial actions decreases. This is because it is difficult
     for outside shareholders, without the inside information and specialized knowledge of managers,
     to ascertain the menu of investment opportunities available to the firm."

Where the principal delegates to an agent the right to initiate and implement decisions, it is important for
the principal to retain decision control authority, i.e., the right to ratify and monitor decisions (e.g.,
Fama and Jensen, 1983; Brickley et al., 1997). Because managers have to make and implement more
investment decisions when growth options are high, we expect that monitoring would increase to
counteract the increase in managerial discretion.3 Oversight by boards of directors, particularly outside
directors, is a common and generally effective form of monitoring (e.g., see Fama and Jensen, 1983).

In addition, we expect monitoring by directors, who represent shareholders' interests, to be a particularly
important form of monitoring in high growth firms because these firms rely more on equity financing than
low growth firms. To be specific, Myers (1977) contends that agency costs of debt will be lower when
firms have more real assets than real options because assets-in-place are generally acquired for specific
business purposes. This constrains managerial discretion and lowers borrowing costs. On the other hand,
real options can lead to underinvestment in the presence of fixed debt because debtholders have a priority
claim to the cash flows arising from positive NPV projects (Myers, 1977). To control for
underinvestment, Myers (1977) predicts that firms with more investment opportunities are likely to use
equity rather than debt financing, and in their studies Smith and Watts (1992) and Gaver and Gaver
(1993) find an inverse relation between IOS and debt/equity ratios.

While agency problems between shareholders and debtholders are minimised when less debt is used, in
high growth firms, conflict between shareholders and managers will increase as equity financing puts
fewer restrictions on management activity relative to covenant-based debt (Skinner, 1993). Because the
value of high growth firms is dependent upon managers' discretionary investment decisions, absentee
shareholders will put in place ex ante mechanisms, such as outside directors, to monitor the initiation and
implementation of decisions by management. Thus, we hypothesise:
H1       Ceteris paribus, the proportion of outside directors on the board will be positively related to the
         level of IOS.

3.   Institutional environment

An important factor which is likely to influence the relation between board composition and IOS is the
institutional environment in which the corporation operates (Smith and Watts, 1992). We argue that the
institutional environment in New Zealand helps to provide a clean and powerful test of the IOS hypothesis
for two reasons.

First, recent changes in New Zealand’s company law have clarified and codified directors' duties. For
example, the Companies Act 1993, which became effective 1 July 1994, requires that directors: 1/ act

                   Another way to address these agency costs is to better align the managers' and
shareholders' interests. For example, both Smith and Watts (1992) and Gaver and Gaver (1993) find a
positive relation between IOS and the incidence of market-based incentive contracts in the US corporate
sector. In New Zealand, however, market-based incentive contracts are relatively rare.
in good faith and in the best interests of the company; 2/ exhibit care, diligence, and skill that a
reasonable person would exercise; 3/ avoid conflicts of interest; and 4/ exercise their powers for
legitimate corporate purposes.

The new Companies Act also increased directors' responsibilities regarding distributions to shareholders,
repayments to creditors, and the accuracy of the company's financial statements. For example, under
section 52 of the Companies Act 1993, directors are required to certify that the company will satisfy a
solvency test before authorising a distribution (i.e., dividends) to shareholders. Likewise, directors now
must ensure that financial statements of the reporting entity comply with applicable financial reporting
standards (i.e., generally accepted accounting practice).

These changes were brought about by problems with the previous companies law, i.e., the Companies
Act 1955. For example, in 1989, the Law Commission noted that "the present [company] law relating
to duties of directors is inaccessible, unclear, and extremely difficult to enforce" (quoted in Hodder,
1993, p. 38). Thus, by specifying directors' duties, the Companies Act 1993 highlights the director's
monitoring role in corporate governance. In addition, because the monitoring function falls on the outside
directors, the Companies Act 1993 sharpens the distinction between inside and outside directors, and this
strengthens the power of our tests.

Second, unlike the US where there have been several proposals to require boards with a majority or
super-majority of outside directors (see Bhagat and Black, 1997, for a review), there have been no similar
proposals in New Zealand. This is important because rather than adding outside directors to reduce
agency costs, US firms may also add outside directors to avoid attention from activist shareholder groups.
More simply, at least some US firms may use outside directors for cosmetic reasons or as “window-
dressing”. Thus, using New Zealand data, we avoid this competing hypothesis and so provide a cleaner
test of the IOS hypothesis.

4. Research design

4.1. Sample selection
The sample used in this study initially comprised 80 firms selected from the 129 companies listed on the
New Zealand Stock Exchange (NZSE) as at 31 December 1995 and included in the Share Market Review
(1995) published by the NZSE. The sample consists of firms that: 1/ responded to our written request
for annual reports for the year 1995; and 2/ agreed to participate in a postal questionnaire survey
requesting information about the equity ownership structure, board composition and characteristics of
board of directors. Ninety-four firms responded to our requests for annual reports. Of the 94 firms, 80
firms agreed to participate in the postal survey. However, this sample was reduced to 77 firms after three
firms were excluded as outliers (see section 5). Thus, our final sample covers approximately 60 percent
of the companies listed on the NZSE in 1995.

Financial data items were extracted primarily from the published annual reports and supplemented with
information obtained from Datex, a NZ financial information provider, and survey responses. While we
examined early and late responses and detected no statistical differences in their responses, if non-
respondents have systematically weaker corporate governance system (e.g., use less outside directors),
the research findings may not be generalisable to all NZ firms. Therefore, we recognise this as a
limitation of our research design.

4.2. Variables
Outside directors are one of several internal mechanisms (e.g., stock option plan, auditing, voluntary
disclosures) that a firm could use to control agency problems and improve firm value. Because the use
of outside directorships may be correlated with other internal monitoring mechanisms, as they are all
driven by a common firm-specific factor - IOS (Rediker and Seth, 1995), corporate governance decisions
are likely to be made simultaneously within the firm. However, Smith and Watts (1992), Gaver and
Gaver (1993) and Barclay and Smith (1995) argue that allowing for interdependencies (simultaneity)
among policy variables would require specification of a system of simultaneous equations. At present the
IOS literature does not provide adequate insights or direction to allow us to identify the appropriate
structural form of this system of equations, and we adopt the approach used in much of the prior
literature, i.e., we examine a single corporate policy decision and assume IOS is pre-determined.4

In line with Smith and Watts (1992), Skinner (1993), Barclay and Smith (1995) and Mian (1996), we
specify the board structure-IOS relation using a reduced form Ordinary Least Squares (OLS) model
where the
proportion of outside directors on the board is the dependent variable and IOS is an independent variable.

                     Titman and Wessels (1988), Jensen et al. (1992), and Agrawal and Knoeber (1996) are
examples of studies that use a simultaneous equation framework. However Smith and Watts (1992, p.
269) also comment that if "the structure they use is correct, the power of their estimates is increased, but
if their structure is incorrect, they impose bias. Given our current knowledge of these empirical relations,
we believe progress is better served by documenting robust empirical relations between policy parameters
and exogenous variables before attempting to subdivide the relations into component effects."
However, because IOS is not the only determinant of board composition, we control for other
determinants of board structure in our OLS model.

4.2.1 Dependent variable
We define the proportion of outside directors on the board (DIR) as the number of independent outside
directors on the board divided by the total number of directors on the board. Independent outside directors
are board members who: 1/ are not an active or retired employee of the firm, and/or 2/ do not have close
business ties with the firm (e.g., a consultant or supplier). While our measure distinguishes between
affiliated outside directors and independent outside directors (e.g., as in Lee et al., 1992), we collect this
data by a postal questionnaire rather than from annual reports or proxy statements as most US studies
have done. We use postal questionnaires because disclosures about directors in New Zealand are limited
and because we wanted to collect information on CEO tenure and the number of board meetings per year
(see section 4.2.3). However, we did check the reliability of the data obtained from the postal survey by
comparing them with data derived from the annual reports for a sub-sample of firms. We detected no
significant discrepancies.5

4.2.2 Independent variable - IOS
IOS is measured by various proxies employed in the prior research. We use four of the most common,
specifically: 1/ market value of the firm to book value of assets (e.g., Smith and Watts, 1992; Barclay
and Smith, 1995; Baber et al., 1996 ); 2/ market to book value of equity (e.g., Gaver and Gaver, 1993;
Lang et al., 1996); 3/ price-earnings (P/E) ratio (e.g., Smith and Watts, 1992; and Gaver and Gaver,
1993); and 4/ an ex-post measure, asset growth (e.g., Pilotte 1992)

The market value of the firm to book value assets, MKTBKA, is a measure of the percentage of firm
value attributable to assets-in-place. This measure assumes firms with more growth options will have
market values far in excess of their book values. MKTBKA is computed as the market value of assets
divided by book value of assets, where market value of assets is defined as the reported value of debt plus
the market value of equity. Gaver and Gaver (1993) point out that using MKTBKA induces bias for firms
with long-lived assets because typically assets are measured on a historical costs basis. However, because
New Zealand-based companies can revalue their non-current assets to show current costs in the financial
statements, this problem is not severe in the present study. Also, MKTBKA is the growth measure used
most frequently in prior studies (e.g., see Jung et al., 1996; Mian, 1996).

                   A copy of the questionnaire is available from the first author on request.
Another measure used in studies such as Gaver and Gaver (1993) and Barclay and Smith (1995) is the
market equity to book equity ratio or MKTBKE. Because high growth firms will have more intangible
assets that are not recorded but are priced by the market, MKTBKE should increase with increases in
growth opportunities. MKTBKE is measured by the ratio of market value of equity to market value of
assets where the market value of equity is estimated by the number of shares outstanding multiplied by
share price at the calendar year end. A problem with using MKTBKE to proxy for IOS is that MKTBKE
could also reflect other factors such as the ability of firms to earn monopoly rents on assets-in-place, and
competition across firms (Ahmed, 1994), and the expected return on equity and risk (Penman, 1996).

The third IOS measure is the P/E ratio (e.g., Chung and Charoenwong, 1991; Gaver and Gaver, 1993).
The rationale for this measure is that because a firm's share price reflects future earnings and not the
current period's income, the difference between current and future earnings will be more pronounced for
high growth firms. As a result, high growth firms will trade at higher P/E ratios than low growth firms.
However, a problem with this measure (as well as MKTBKA and MKTBKE) is that it relies on the
market value of corporate shares. Because share prices decrease with increases in leverage, the market-
based measures of IOS are likely to be sensitive to the firm's capital structure (Gaver and Gaver, 1993).

The fourth measure is the growth in assets or ∆ASSETS. Pilotte (1992) contends that, under rational
expectations, actual subsequent growth should be good proxy for anticipated investment. One problem
with the ex post proxy is that they measure actual growth rather than the profitability on new investment.

While these are the most common IOS measures, because each has limitations, we also factor analyse
to construct a composite measure of IOS.6

4.2.3. Control variables

                    Because most firms in our sample are not engaged in research and development (R&D),
we omit this variable even though it has been used in prior studies (e.g., Gaver and Gaver, 1993).
Industry membership is another variable that has been used as a proxy for IOS in prior research (e.g.,
Chan et al., 1990). However, as Gaver and Gaver (1993) argue, cross-sectional variation in IOS is also
likely at the firm level. Because we are interested in board composition which is a firm-specific decision,
we also do not use industry membership.
In testing H1, we control for other factors that may also affect board composition. For example, the
degree of incentive conflicts and agency costs will be a function of managerial ownership (e.g., Jensen
and Meckling, 1976; Fama and Jensen, 1983). Firms with low levels of managerial ownership will have
more demand for outside directors, and prior studies (e.g., Bathala and Rao, 1995; Rediker and Seth,
1995) empirically document such a relation.

We include inside ownership (defined as a percentage of common equity owned by directors and the “top
five” managers) as a control variable. We measure inside ownership using a dummy variable. Morck et
al., (1988) argue that firms that are 5 to 25 percent owned by management are likely to have lower
market values. Consequently, we use a dichotomous measure INSDUM as a surrogate for inside
ownership. INSDUM is coded 1 if inside ownership is less than or equal to 5 percent and greater than
or equal to 25 percent , and 0 elsewhere.

Likewise, Jensen and Meckling (1976) suggest that agency conflicts between shareholders and debtholders
increase with leverage. To the extent that outside directors can mitigate these conflicts, the demand for
outside directors should be positively related to levels of debt. In addition, because debt imposes fixed
costs on the firm and increases the possibility and costs associated with bankruptcy (Jensen and Meckling,
1976), the need for additional monitoring by outside directors will also increase. We include leverage as
a control variable where leverage is defined as the book value of long term liabilities divided by total
assets (LEV).

Hermalin and Weisbach (1988, 1991) suggest that long serving CEOs have a greater influence over the
selection of board members and may include directors who are likely to be closely aligned with the
CEO's interests. Menon and Williams (1993) argue that the effectiveness of outside directors will
increase when boards meet more frequently. This suggests that outside directors are more useful when
meetings are frequent, and meeting frequency should be positively related to the proportion of outside
directors. In any case, we include CEO tenure (CEOTEN) and the number of board meetings as
additional controls. Both items were obtained from our postal questionnaire survey.

Firm size has been included as a control variable because prior research (e.g., Garver and Garver, 1993;
Mehran, 1995) suggests that firm size is related to incentive control mechanisms including debt, dividends
and managerial compensation. Further, Gaver and Gaver (1993) and Barclay and Smith (1995) predict
that large firms have more information asymmetry between managers and outside stakeholders, and this
creates demand for outside directors. However, Rosenstein and Wyatt (1990) suggest small firms will
have more outside directors because large firms can rely on alternative monitoring mechanisms (e.g.,
institutional investors, stock analysts). Consistent with previous IOS studies (e.g., Mian, 1996), we use
the book value of assets minus the book value of common equity plus the market value of common equity
as a proxy for firm size. This amount is transformed using a natural log and is labeled LNSIZE.
However, given the prior discussion, no directional sign is predicted for the coefficient.

4.2.4    Model specification
We test H1 using the following OLS model:
(1)      DIR = b0 + b1 IOS + b2 INSDUM + b3 LEV + b4 CEOTEN + b5 BRDMEET + b6 LNSIZE
where DIR, INSDUM, LEV, CEOTEN and LNSIZE are those described in sections 4.2.1-4.2.3, and
b0 is the intercept and e is an error term which is assumed to be normally distributed with zero mean.
IOS is either MKTBKA, MKTBKE, P/E, GSALES or a composite IOS measure depending on the exact

5. Results

5.1. Main statistical tests
Table 1 provides descriptive statistics for DIR and the right-hand side variables. We also examined scatter
plots of the distribution for the variables and identified two extreme values for P/E (i.e., 110 and 73.3)
and extreme values for MKTBKA (9.314) and MKTBKA (19.01) for another firm. Because each of these
values are more than 59 percent larger than the next highest value, these three firms are deleted, reducing
the sample to 77 from 80.

The mean for DIR is 52 percent indicating that on average outside directors comprise a majority of the
board. The mean of inside ownership is 17 percent, and the mean for CEOTEN is 6 years. Each of these
values is similar to those reported in prior US-based studies (e.g., Mehran, 1995; Dechow et al., 1996).

                                        <Insert Table 1 here>

Table 2 gives a correlation matrix for the IOS and control variables. As predicted, individual IOS
measures such as MKTBKA and MKTBKE are significantly and negatively correlated with leverage (
r = -0.409, p ≤ 0.001 and r = -0.379, p ≤ 0.001 respectively) which is consistent with previous IOS
research (e.g., Smith and Watts, 1992; Gaver and Gaver, 1993; and Barclay and Smith, 1995 ). The
INSDUM variable is significantly and negatively correlated with firm size ( r = -0.349, p ≤ 0.01) which
is similar to prior studies (e.g., Jensen et al., 1992). Also, we find a significant positive correlation
between firm size and P/E ratio (r = 0.429; p ≤ 0.001).

Regarding correlations between individual IOS measures, only MKTBKA and MKTBKE are positively
and significantly correlated (r = 0.911, p ≤ 0.001). However, there are no significant correlations
between other IOS variables. This is surprising as Gaver and Gaver (1993) found statistically significant
correlations among all individual IOS variables. However, our correlation tests are not nearly as powerful
as the ones used by Gaver and Gaver (1993) since they use 1525 US firm-based observations.

                                         <Insert Table 2 here>

Table 3 provides the multiple regression results for equation 1. The four models shown differ in term of
the proxy used for IOS. Each of the models is statistically significant with R2 varying between 20 percent
and 26 percent. The R2 statistics are comparable to similar models using US data (e.g., Brickley and
James, 1987, R2 = 26 percent; Bathala and Rao, 1995, R2 = 19 percent). Though all four IOS variables
have positive signs, only two, MKTBKA and MKTBKE, are statistically significant (p = 0.01 based on
one-tailed tests). Thus, we find some evidence that investment opportunities require additional monitoring
and give managers incentives to hire independent outside directors (Gaver and Gaver, 1993; Booth and
Deli, 1996). This supports H1.

                                         <Insert Table 3 here>

Of the control variables, firm size is significant at 0.05 level or better (two-tailed tests) with a negative
sign in all three models. This is consistent with Rosenstein and Wyatt's (1990) view that small firms have
fewer alternative monitoring mechanisms and, consequently, will rely on outside directors. INSDUM is
significant with a negative sign in each of the models indicating that at low levels of managerial
ownership, inside ownership is not sufficient to align the manager's interests with the shareholder's
interests, and that at these levels, management may try to entrench itself by adding inside or affiliated
outside directors.

The coefficient for BRDMEET is positive and significant at 0.05 level (one-tailed tests) in all four
models, indicating that frequently meeting boards are more likely to include outside directors to protect
against such behaviour as earnings management and misuse of free cash flows (Menon and Williams,
1993). LEV is positively related to DIR at the 0.10 level (one-tailed test) in models 1 and 2 which
suggests that firms with more debt rely on outside directors to minimise conflict between shareholders
and debtholders. Finally, CEOTEN is not statistically related to board composition in any of the models,
suggesting that this variable is not related to board composition which is contrary to US-based results
reported by Hermalin and Weisbach (1988) and Bathala and Rao (1995).7

                 Following Hermalin and Weisbach (1991), the regressions were re-run using four
piecewise variables for CEOTEN. However, the results were unchanged, and so these results are not
reported here.
5.2. Additional tests8

5.2.1 Sensitivity analysis
The empirical results presented in the preceding section select a particular point in time to observe the
relation between board composition and IOS. One problem with the cross-sectional tests is that they do
not examine explicitly whether the fraction of outside directors presently employed by a firm is due to
the firm’s current IOS or due to its historical IOS. Furthermore, the cross-sectional tests do not examine
the endogeneity, or reverse causation, between the fraction outside directors and IOS. For example, it
is possible that a higher proportion of outside directors provides more effective monitoring, and this leads
to higher company performance as measured by market to book ratios.9 Finally, our coefficient estimates
may be “contaminated” by time invariant fixed firm effects.

To partially address these problems, we investigate whether changes in the fraction of outside directors
are related to the changes in IOS on a sub-sample of 53 firms10. Changes in the fraction of outside
directors and IOS in year t are computed for each firm i over a five year period 1991-95. For example,
the change in the fraction of outside directors in year t is computed by the ratio of an increase (decrease)
in the fraction of outside directors from t to t-4 divided by the fraction of outside directors for the year
t-4 where t-4 is 1991 and t is 1995. Thus, we estimate the following changes model:

(2)      ∆DIR =b0 + b1 ∆IOS + ∆e
where b1 represents an uncontaminated estimate of the relation between IOS and DIR (i.e., fixed effects
removed) and where b0 can be interpreted as a time effect.

Table 4 contains the results for the changes model. The results indicate that two of the three IOS models
are significant. Contrary to expectations, the coefficient for ∆P/E is not significant at conventional levels.
However, the coefficients on ∆MKTBKA and ∆MKTBKE are positively and significantly related at 0.01

                   As in Gaver and Gaver (1993), we used factor analysis to extract a composite IOS
measure form the individual IOS variables. The composite measures (proxied by IOSFAC and
IOSDUM) yielded results similar to those reported in Table 3. Both IOSFAC and IOSDUM were
significant at 0.05 level. To economise on space, however, we do not present these results in this
                   We are grateful to an anonymous reviewer for making this point. Theoretically, a two-
stage least square (2SLS) should be used to test simultaneous effect between the DIR and IOS. Because
IOS hypothesis is not adequately developed and because a simultaneous relationship between DIR and
IOS is difficult to predict on an ex ante basis (Bhagat and Black, 1997), 2SLS is not estimated in this

         10       Estimation of a fixed-effects covariance model proved difficult on the small cross-
sectional and time series data set available. In particular, there was collinearity between the dummy
variable INSDUM and the fixed-effects parameters that inhibited the derivation of complete and
meaningful coefficient estimates.
and 0.05 levels respectively. These results are consistent with our cross-sectional results and suggest that
the fraction of outside directors increases with contemporaneous increases in IOS. Taken together with
our earlier tests, the results for IOS appear to be robust.

                                         <Insert Table 4 here>

6. Summary and conclusions

This study has examined empirically the determinants of board composition in the NZ corporate sector.
Based on Smith and Watts (1992) and Gaver and Gaver (1993), we use an efficient contracting
perspective and predict that the use of independent outside directors will differ between high growth and
low growth firms. The empirical results obtained from cross-sectional tests of 77 New Zealand firms
indicates that the proportion of outside directors is significantly and positively related to IOS, leverage,
and the number of board meetings, and is significantly and negatively related to low levels of inside
ownership and firm size. In contrast, CEO tenure and high levels of inside ownership do not appear to
be significantly related to the proportion of outside directors, at least for our sample. Our results support
the view that high growth and the use of outside directors are related. Moreover, we show that the change
in outside directors is related to contemporaneous changes in IOS.

While both the determinants of board composition and the effect of IOS on corporate policy decisions
have been examined extensively in the literature, our study is the first to explicitly examine the effect of
IOS on board composition. In doing so, we contribute to, and provide a link between, the largely separate
IOS and board composition literatures. From a policy standpoint, our results also suggest that the optimal
composition of the board, in terms of inside and outside directors, will not be the same for high and low
firms. Consequently, mandating that all firms use a majority or super-majority of independent outside
directors could result in less efficient (i.e., sub-optimal) board structures from a societial point of view
(Jensen, 1993).

As with any research, a few caveats deserve mention. First, the empirical tests carried out in our study
may suffer from omitted variables. For instance, theoretical and empirical research suggests the number
of outside directorships held by CEOs is likely to be affected by supply side factors such as the CEO's
career cycle and interlocking boards (Booth and Deli, 1996). Therefore, it is possible that the empirical
relations shown in our study could be driven by cross-sectional differences in omitted supply-side factors
that affect board composition. A second limitation arises from possible data measurement errors. If our
proxies for IOS and outside directors are noisy, measurement error could potentially reduce the statistical
power of the tests. Third, the explanatory variables used in our study are considered exogenous, but
Smith and Watts (1992), Gaver and Gaver (1993) and others recognise that several explanatory variables
used in this research, including IOS, may be endogenous. Fourth, the present study is essentially a single
period, small sample analysis. As such, the results reported here might not be as robust as those derived
from studies using large cross-sectional and time series data. Lastly, we examine the efficiency of a single
mechanism in controlling agency conflicts without considering several alternative means by which a firm
can monitor the manager's activity. For instance, a firm can use stock option plans, an audit committee,
external auditors or a mix of these control mechanisms to attain an optimal level of monitoring. We
suggest that each of these limitations provides an area for future research.

Agrawal, A. and Knoeber, C.R. (1996).'Firm Performance and Mechanisms to Control Agency
       Problems between Managers and Shareholders', Journal of Financial and Quantitative Analysis,
       31(September): 377-397.

Alchian, A.A. and Woodward, S. (1986). 'An Evolving Model of Firms'. Journal of Institutional and
        Theoretical Economics, 142: 34-49.

Ahmed, A.S. (1994). 'Accounting Earnings and Future Economic Rents: An Empirical Analysis.'
       Journal of Accounting and Economics, 21(November): 377-400.

Baber, W.R., Janakiraman, S.N., and Kang, S-H.(1994).'Investment Opportunities and the Structure of
        Executive Compensation.' Journal of Accounting and Economics, 21(November): 297-318.

Barclay, M.J. and Smith, C.W. (1995).'The Maturity Structure of Corporate Debt.' The Journal of
        Finance, 50(March): 609-631.

Bathala, C.T. and Rao, R.P. (1995). 'The Determinants of Board Composition: An Agency Theory
         Perspective'. Managerial and Decision Economics, 16(January/February): 59-69.

Bhagat, S. and Black, B. (1997).'The Uncertain Relationship Between Board Composition and Firm
        Performance'. Columbia Law School Working Paper, New York.

Booth, J.R., and Deli, D.N. (1996).'Factors Affecting the Number of Outside Directorships held by
        CEOs.' Journal of Financial Economics, 40(January): 81-104.

Brickley, J.A. and James, C.M. (1987).'The Takeover Market, Corporate Board Composition, and
         Ownership Structure: The Case of Banking.' Journal of Law and Economics, 30(June): 161-180.

Brickley, J.A., Coles, J.L., and Terry, R.L. (1994).'Outside Directors and the Adoption of Poison Pills.'
         Journal of Financial Economics, 35(3), 371-390.

Brickley. J.A., Smith, C.W. and Zimmerman, J.L. (1997). Managerial Economics and Organizational
         Architecture. Chicago: Irwin.

Byrd, J.W. and Hickman, K.A. (1992).'Do Outside Directors Monitor Managers?: Evidence from
        Tender Offer Bids.' Journal of Financial Economics, 32(October): 195-221.

Chan, S.H., Martin, J. and Kensinger, J. (1990).'Corporate Research and Development Expenditures
        and Share Value.' Journal of Financial Economics, 26(August): 255-276.

Chung, K.H. and Charoenwong, C. (1991). 'Investment Options, Assets in Place, and the Risk of
       Stocks'. Financial Management, 20(Autumn): 21-33.

Dechow, P.M., Sloan, R.G. and Sweeney, A.P. (1996).'Causes and Consequences of Earnings
       Manipulation: An Analysis of Firms Subject to Enforcement Actions by the SEC.'Contemporary
       Accounting Research, 13(Spring): 1-36.

Fama, E. and Jensen, M.C. (1983). 'Separation of Ownership and Control.'Journal of Law and
       Economics,26(June): 301-325.
Gaver, J.J. and Gaver, K.M. (1993).'Additional Evidence on the Association between the Investment
        Opportunity Set and Corporate Financing, Dividend, and Compensation Policies.'Journal of
        Accounting and Economics, 16(January/April/July): 127-160.

Hermalin, B.E. and Weisbach, M.S. (1988). 'The Determinants of Board Composition.' Rand Journal
        of Economics, 19(Winter): 589-606.

Hermalin, B.E. and Weisbach, M.S. (1991).'The Effects of Board Composition and Direct Incentives
        on Firm Performance.' Financial Management, 20(Winter): 101-112.

Hodder, J. (1993). 'The bill returns - Directors and liquidations.' Accountants' Journal, 72 (February):

Jensen, G.R., Solberg, D.P. and Zorn, T.S. (1992). 'Simultaneous Determination of Inside Ownership,
         Debt and Dividend Policies.' Journal of Financial and Quantitative Analysis, 27(June): 247-263.

Jensen, M.C. (1986).'Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers.' American
        Economic Review, 76(May): 323-329.

Jensen, M.C. (1993).'The Modern Industrial Revolution, Exit, and the Failure of Internal Control
        Systems.' The Journal of Finance,48(July): 831-880.

Jensen, M.C. and Meckling, W.H. (1976).'Theory of the Firm: Management Behavior, Agency Costs
        and Ownership Structure.'Journal of Financial Economics, 3(October): 305-360.

Jensen, M.C. and Meckling, W.H. (1995). 'Specific and General Knowledge, and Organizational
        Structure'. Journal of Applied Corporate Finance, 8(Summer): 4-18.

Jung, K., Kim, Y-C., Stulz, R.E. (1996).'Timing, Investment Opportunities, Managerial Discretion, and
         the Security Issue Decision.' Journal of Financial Economics, 42(October):159-185.

Klein, A. (1998). 'Firm Performance and Board Committee Structure.' The Journal of Law and
        Economics, 41(April): 275-303.

Lang, L., Ofek, E. and Stulz, R.M. (1996).'Leverage, Investment, and Firm Growth.' Journal of
        Financial Economics, 40(January): 3-29.

Lee, C.I., Rosenstein, S., Rangan, N. and Davidson, W.N. (1992). 'Board Composition and Shareholder
         Wealth: The Case of Management Buyouts'. Financial Management, 21(Spring): 58-72.

Mehran, H. (1995).'Executive Compensation Structure, Ownership, and Firm Performance.' Journal
       of Financial Economics, 38(June): 163-184.

Menon, K. and Willaims, J.D. (1994).'The Use of Audit Committees for Monitoring.' Journal of
       Accounting and Public Policy, 13(Summer): 121-139.

Mian, S.L. (1996). 'Evidence on Corporate Hedging Policy.' Journal of Financial and Quantitative
        Analysis, 31(September): 419-439.

Morck, R., Shleifer, A. and Vishny, R.W. (1988).'Management Ownership and Market Valuation: An
       Empirical Analysis.'Journal of Financial Economics, 20(January/March): 293-315.

Myers, S.C. (1977). 'Determinants of Corporate Borrowing.' Journal of Financial Economics, 5
       (November): 141-175.

New Zealand Stock Exchange (NZSE)(1995), Share Market Review. Wellington: NZSE Press.
Patton, A. and Baker, J. (1987). 'Why won't Directors Rock the Boat?' Harvard Business Review,
        65(September): 10-12.

Penman, S.H. (1996).'The Articulation of Price-Earnings Ratios and Market to Book Ratios and the
       Evaluation of Growth.' Journal of Accounting Research, 34(Autumn): 235-259.

Pilotte, E. (1992). ‘Growth Opportunities and the Stock Price Response to New Financing’. Journal of
          Business, 65(July): 371-394.

Rediker, K.J. and Seth, A. (1995).'Board of Directors and Substitution Effects of Alternative Governance
         Mechanisms.' Strategic Management Journal, 16(February): 85-99.

Rosentein, S., and Wyatt, J.G. (1990).'Outside Directors, Board Independence, and Shareholder
        Wealth.' Journal of Financial Economics, 26(August): 175-191.

Shleifer, A. and Vishy, R. (1997). 'A Survey of Corporate Governance'. Journal of Finance, 52(June):

Skinner, D.J. (1993). 'The Investment Opportunity Set and Accounting Procedure Choice - Preliminary
         Evidence.' Journal of Accounting and Economics, 16(October): 407-445.

Smith, C.W. and Watts, R.L. (1992).'The Investment Opportunity Set and Corporate Financing,
        Dividend, and Compensation Policies.' Journal of Financial Economics, 32(December): 263-

Titman, S. and Wessels, R. (1988). 'The Determinants of Capital Structure Choice.' The Journal of
                                            TABLE 1. Descriptive statistics

       Variablea                    Mean         Median       Std. dev.      Minimum    Maximum       Numberb

       DIR                          0.520         0.500          0.273         0.000       1.000            77

       IOS variables
       MKTBK                        1.425         1.115          0.927          0.542      5.846            77
       MKTBKE                       1.611         1.270          1.264          0.318      7.881            77
       P/E                         11.022         9.950          7.516        -15.790     35.090            77
       ∆ASSETS                      0.185         0.044          0.669         -0.845      3.424            77

       Control variables
       INSDUM                       0.208         0.000          0.408         0.000       1.000            77
       LEV                          0.198         0.161          0.202         0.000       0.775            77
       CEOTEN                       6.003         5.000          4.188         0.500      18.000            77
       BRDMEET                       9.61        11.000          2.987         1.000      14.000            77
       LNSIZE                       7.919         7.940          0.809         0.606      10.132            77

    DIR = number of independent outside directors/total directors; MKTBKA = market value of firm value to book
value of assets; MKTBKE = market value of firm to book value of equity shares; P/E = price per share to
primary earnings per share; ∆ASSETS= Growth in total assets between 1991-95; INSDUM = percentage of
    insider (directors and top five managers) ownership INSDUM is coded 1 if inside ownership is less than or equal
to 5 percent and greater than or equal to 25 percent , and 0 elsewhere; LEV = long term liabilities/total assets;
CEOTEN = number of years the CEO has been on the job; BRDMEET =number of board meetings per year;
LNSIZE = natural logarithm of firm value.

          The sample size of 77 is obtained after omitting three outliers.
                                        TABLE 2. Pairwise correlation coefficients
                 MKTBK          MKTBKE           P/E       ∆ASSETS      INSDUM          LEV      CEOTEN   BRD     LNSIZE
                   A                                                                                      MEET


MKTBKA             1.000
MKTBKE             0.911b         1.000
P/E                -0.111         -0.005        1.000

GSALES              0.397         0.243         -0.172       1.000


INSD               -0.067      -0.098            -0.072       -0.085         1.000
LEV                -0.409b     -0.379b           0.033        0.186           0.038     1.000
CEOTEN             0.053        0.056           -0.083         0.044          0.194     -0.120   1.000
BRDMEE             0.005        0.007            -0.133        0.078          -0.147     0.182   0.002

LNSIZE           -0.078        0.143            0.429b         0.318c         -0.349c    0.050    0.086   1.000

    See Table 1 for variable definition and sample size.
    Significant at 0.001 level based on two-tailed test.
    Significant at 0.01 level based on two-tailed test.
                                                 TABLE 3. Estimated Coefficients (t-statistic in parentheses) from OLS Regressiona,b

                      Individual IOS Variables                                           Control Variables
Mdl           MKTBKA          MKTBKE          P/E          ∆ASSETS       INSDUM             LEV        CEOTEN        BRD         LNSIZE       Incpt       MPrb        R2
                +               +              +              +             -                +            -          MEET         +/-          +/-
    1           0.084                                                     -5.801            0.388        0.004       0.017        -0.103      1.163       0.005     26.2%
               (2.565)c                                                  (-2.002)c         (1.516)e     (0.635)     (1.754)d     (-2.299)d   (3.064)
    2                          0.060                                      -5.377            0.398        0.005       0.017        -0.124      1.344       0.005     25.9%
                              (2.505)c                                   (-1.869)c         (1.541)e     (0.656)     (1.759)d     (-2.669)c   (3.490)
    3                                         0.004                       -4.816            0.181        0.005        0.019        -0.107       1.251      0.044     19.8%
                                             (0.915)                     (-1.662)c         (0.701)      (0.589)      (1.877)d     (-2.097)d    (3.083)
    4                                                                0.406             -4.873             0.330          0.062          0.018       -0.094        1.355
            0.026     19.3%                                                              (0.487)           (-1.920)c          (1.659) e     (0.697)     (1.753) d (-2.037)d

        Dependent variable = DIR. See Table 1 for variable definitions. Sample sizes are 79, 79, 78 , and 80 after omitting outliers.
        Model probability is based on a F-statistic. Significance levels for t-statistics are based on one-tailed tests except for LNSIZE.
        Significant at 0.01 level.
        Significant at 0.05 level.
        Significant at 0.10 level.
TABLE 4. Estimated coefficient (t-statistic in parentheses) from OLS regression changes modela

                                 ∆ IOS variables
  Mdl             ∆MKTBKA          ∆MKTBKE           ∆P/E         ∆ASSETS     Incpt       MPrbb           R2
                     +                +               +               +        +/-

       1             0.110                                                     0.404       0.005        14.34%
                    (2.922)c                                                 (10.016)
       2                              0.054                                    0.422       0.020        10.87%
                                     (2.494)d                                (10.690)
       3                                                                      0.468        0.596         0.56
                                                      ( -0.534)

                                                                   -0.036     0.127
   4                                                                                       0.058         0.025
                                                                  (-0.536)   (1.938)

     Dependent variable = ∆DIR where ∆DIR = DIR1995 - DIR1991/DIR1991. ∆MKTBKA = MKTBKA1995- MKTBKA1991/MKTBKA1991; ∆MKTBKE = MKTBKE1995-
    MKTBKE1991/MKTBKE1991; ∆P/E = P/E1995 -P/E1991/P/E1991, ∆Assets = ASSETS1995 – ASSETS1991/ ASSETS1991. Sample size is 53 based on all firms with data available
   for 1991 and 1995.
           Model probability is based on a F-statistic.
       Significant at 0.01 level based on one-tailed test.
           Significant at 0.05 level based on one-tailed test.

To top