Differences in Agency Problems between Public and Private Firms: Evidence from Top
Management Turnover
Ugur Lel
Virginia Tech
[email protected] Darius Miller
Southern Methodist University
[email protected] Natalia Reisel
Fordham University
[email protected] Abstract
We compare a primary outcome of corporate governance, the propensity to replace poorly
performing managers, between public and private firms in a large cross-country sample. We
show that public firms display a higher sensitivity of top management turnover to firm
performance than private firms. We find that this difference in the sensitivity of management
turnover to firm performance stems from the information production and monitoring function of
public equity markets and from the market for corporate control. Our results suggest that agency
problems in publicly traded firms may be less severe than previously anticipated and financial
markets play an important governance role.
_______________ We would like to thank participants of the Early Ideas Session at the 2011 Financial Research Association
Meeting and the 2013 Conference on Empirical Legal Studies at the University of Pennsylvania. We also
would like to thank Ulrike Shultze for helping us classify some of the legal forms not on the Bureau van
Dijk list.
1 How severe are the agency problems in public corporations? The goal of this paper is to address
this important question by comparing public corporations’ propensity to dismiss poorly
performing managers to their natural benchmark: private firms. Further, we seek to provide
evidence into the governance mechanisms that may drive the differences in agency problems
between public and private firms.
Since Berle and Means (1932), the separation of ownership and control in public
companies has been argued to cause potentially severe agency problems between managers and
shareholders, including perquisite consumption and empire building. Moreover, managers of
public corporations who engage in such value destroying actions may do so with impunity since
public companies are also characterized by disperse ownership structures that can lead to
managerial entrenchment. However, incorporating as a public company is also often argued to
promote good corporate governance since listed companies benefit from the scrutiny of the
public equity markets and the market for corporate control (see, e.g., Manne (1965), Jensen (
1993), Holmstrom and Tirole (1993), Dow and Gorton (1997)), which may lead to less severe
agency problems in public than private firms.
Despite the importance of knowing the relative severity of agency problems between
public and private firms, evidence has been limited by the fact that public firms' natural
benchmark, private firms, are not required to file financial reports in the U.S. This lack of
disclosure by private U.S. companies has forced researchers to draw inferences from limited
samples of private firms, such as firms from single industries (Sheen (2009)), large private firms
with access to public debt markets (Coles, Lemmon, and Naveen (2003)) and private companies’
non-financial data, such as corporate jet usage (Edgerton (2012)).
2 In this paper, we exploit the fact that in the European Union, private firms face similar
reporting requirements as public firms. This extensive disclosure requirement allows us to
benchmark public firms with a comprehensive group of private firms. We are able to compare a
direct outcome of corporate governance, the propensity to replace value-destroying managers,
across thousands of public and private corporations to assess the quality of governance and the
severity of managerial agency problems in public and private firms. Replacing poorly
performing CEOs is argued to be a necessary condition for good corporate governance (Shleifer
and Vishny (1989, 1997)), and the sensitivity of top executive turnover to firm performance as a
measure of the quality of corporate governance has been supported by a large number of studies
in the U.S. and abroad, including recent research by Dahya, McConnell, and Travlos (2002),
Volpin (2002), Gibson (2003), DeFond and Hung (2004), and Aggarwal, Erel, Ferreira, and
Matos (2011)).1 Furthermore, since our data span several countries, we are also able to
investigate the governance mechanisms that drive the differences in agency problems between
public and private firms by examining how cross-country differences in the market for corporate
control and scrutiny of public equity markets affect top management turnover.
We find that publicly traded firms display a higher sensitivity of top management
turnover to poor firm performance than both private as well as unlisted public firms.2 This result
survives a battery of robustness tests including alternative top management turnover measures
and alternative sample selection procedures, including matched and unmatched samples of listed
and unlisted firms. We also consider a number of alternative firm performance measures and find
no evidence that unlisted firms are more likely to replace poorly performing managers than listed
1
For U.S.-based studies, see Hermalin and Weisbach (2003) and citations contained therein.
A public firm in our sample is defined as listed if its stock is traded on a major exchange; otherwise, the firm is
defined as unlisted. We use the terms private and unlisted interchangeably to refer to both private firms and unlisted
public firms, and listed and public to refer to publicly traded firms.
2
3 firms. Overall, our findings suggest that public firms are more likely to replace value-destroying
managers than private firms and therefore are inconsistent with the view that managers of public
companies are more likely to be entrenched than managers of private companies.
We next provide evidence on the potential mechanisms that may contribute to the higher
sensitivity of management turnover to poor firm performance in public corporations. One
potential mechanism that facilitates the replacement of poorly performing managers in public
firms is the market for corporate control. For example, Manne (1965) and Jensen (1993)
emphasize the importance of the market for corporate control in disciplining managers in public
corporations. Exploiting cross-country differences in the extent of anti-takeover regulations, we
find that listed firms are more likely to replace poorly performing managers than unlisted firms
in countries that limit antitakeover tactics.
Another potentially important governance mechanism for publicly traded companies is
the scrutiny of public equity markets. The information production and monitoring role of stock
markets in alleviating agency problems in public corporations has been emphasized in a number
of studies including Fama (1980), Holmstrom and Tirole (1993), Dow and Gorton (1997),
Subrahmanyam and Titman (1999), Gupta (2005), Edmans (2009). Again exploiting the crosscountry nature of our data, we investigate if differences in the availability of firm-specific
information in public equity markets (proxied by stock market synchronicity) are related to the
higher sensitivity of management turnover in listed firms. We find that listed firms are more
likely to replace poorly performing managers than unlisted firms only in countries where stock
markets produce more firm specific information. Taken together, our results on two important
mechanisms of public incorporation provide evidence that financial markets are not just a
4 sideshow but rather an important channel that may limit managerial entrenchment and mitigate
agency problems in public corporations.
We investigate several alternative explanations for our results. One plausible explanation
is that our findings could be driven by already well-governed firms self-selecting to become
listed. We test this alternative explanation directly by exploiting the sub-sample of firms that
change status from listed to unlisted or vice versa. If only well-governed firms select to be listed
and the change in status itself doesn’t affect corporate governance, then there should be no
difference in the sensitivity of top management turnover to firm performance before and after the
change in status. In contrast, we find that the change in status from unlisted to listed is associated
with an increase in sensitivity of CEO turnover to performance, which further confirms the
potential benefits associated with public equity markets. Further, to mitigate potential selfselection biases, we use a matched sample of listed and unlisted firms throughout our analysis.
We also investigate whether our results are explained by the relatively high managerial
ownership found in private firms since higher inside ownership may lead to managerial
entrenchment. After excluding unlisted firms that are likely to have high managerial ownership,
we continue to find that listed firms display higher sensitivity of top management turnover to
firm performance than unlisted firms. Moreover, we investigate how family-run firms may
impact the sensitivity of top management turnover in public and private firms. We find that the
high sensitivity in listed firms is concentrated in non-family-run firms and that the sensitivity of
top management turnover to performance in family-run listed firms is not different from the
sensitivity in unlisted firms. These findings are consistent with Villalonga and Amit (2006) who
find that family-run public firms trade at a significant discount compared to non-family public
firms, suggesting that agency conflicts are quite severe in family-run public firms. They are also
5 in line with Volpin (2002) who shows that family-run firms have weaker sensitivity of CEO
turnover to poor firm performance than other firms in the Italian stock exchange.
Finally, we examine whether the lower sensitivity of top management turnover to poor
firm performance in private companies is driven by private firms’ ability to consider longer
performance windows in evaluating managers’ performance. For example, Boot, Gopalan and
Thakor (2006, 2008) suggest a benefit of private ownership is that it enables the manager to
achieve optimal level of decision-making discretion (or autonomy) through private contacting
with a few large investors, while such discretion is not visible in public firm because of
constantly changing investor base. We perform a series of tests where performance measures are
lagged up to three years and continue to find that listed firms are significantly more likely to
replace poorly performing top management.
Our findings make several contributions to the literature that examines agency problems
in public corporations. First, our results suggest that some agency problems may not be as severe
as previously thought. For example, a closely related to our paper is the recent study by Edgerton
(2012) who investigates agency problems in public corporations by comparing the use of
corporate jets across public and private firms. He finds evidence that, in at least a subset of
public firms, executives engage in excessive perquisite consumption. Our results, however, show
that such value destroying activity in public firms is met with more severe consequences:
executives of public firms that engage in value destroying activities are more likely to be
replaced than value destroying managers of private firms.
Second, our paper also contributes to the literature on the real effects of financial
markets. For example, Edmans, Goldstein, and Jiang (2012) identify empirically a strong effect
of market prices on takeover activity, thus showing that the market is not a sideshow but rather
6 exerts a powerful disciplinary effect on management. We add to this literature by demonstrating
that public firms are more likely to replace poorly performing managers than private firms in
countries where anti-takeover provisions are weak.
More broadly, our results contribute to the growing literature that examines various
aspects of corporate policy across private and public firms using a comprehensive set of public
and private firms. Giannetti (2003), for instance, compares the capital structure of listed and
unlisted European firms. Michaely and Roberts (2012) investigate the dividend policy of public
and private firms in the UK.
The remainder of the paper proceeds as follows. Section 1 details our sample selection
procedure and describes the data. Section 2 present results of the top management turnover
analysis, including cross-country tests, alternative explanations and robustness. Section 3
concludes.
1. Data and Descriptive Statistics
1.1. Sample Selection and Variables
Our primary data source is the 2011 version of Amadeus provided by Bureau van Dijk. This
database provides balance sheet and income statement items as well as information on company
managers for a large set of European firms. An important advantage of Amadeus is that it
includes data for public and private firms, which is made possible in part because European laws
require both public and private firms to report financial statements. The data are collected from
each national official public body in charge of collecting the annual accounts in its country, and
always come from the officially filed and audited accounts.
7 We classify firms into listed and unlisted based on the data field listed in Amadeus.
Listed firms are those with a legal status of public in Amadeus and have their shares traded on a
main stock exchange in the country. The rest, public firms whose shares are not publicly traded
on a main exchange and private firms, are labeled as unlisted firms. We use the field legal form
to identify private firms and exclude unlimited partnerships, sole proprietorships, cooperatives,
foreign companies, foundations, and government enterprises.3 In our analysis, we combine
private firms and unlisted public firms in one group because both types are not exposed to
benefits associated with access to active public equity markets, and thus we focus on differences
between listed and unlisted firms.4 We use data from the historical Amadeus DVDs to track the
listing status over time. Firms can stay publicly traded throughout the sample period or become
unlisted (i.e., go private) and vice versa over time.
The data on managers for private and public firms are obtained from historical Amadeus
DVDs. For top executives, the dataset reports names and positions within a company starting in
1999. For many private companies in Europe, the firm’s top managers are not classified as chief
executive officers but instead are most often reported as directors, managing partners, or
managing directors. Given the difficulty in identifying the top executive in a firm, we follow
previous research that computes turnover statistics for the entire top management team.5 For
example, Mikkelson and Partch (1997) measure management turnover as a change if the CEO,
president, and the chairman are replaced in their time period.
After we identify the top management team in each firm, we create an alias for each
manager using the first letter of first names and the last names, and use this variable to track
3
Bureau van Dijk provides a list to classify legal forms within each country into public and private organizational
types.
4
Also, Mortal and Reisel (2013) show that unlisted public firm behave similarly to private firms.
5
We identify a set of top management titles for each country by using top manager titles in Gibson (2003), Defond
and Hung (2004), and Lel and Miller (2007) and supplementing these sources through online searches.
8 managers over time within each firm in our calculation of turnover measures. In some instances
managers' names or titles are reported in Amadeus together with their affiliation with other firms.
We drop such affiliations in our creation of managerial aliases.
We create an indicator variable to measure a turnover event which takes on the value of
one whenever at least half of the top management team is turned over (turnover dummy I). While
the 50 percent cutoff point is arbitrary, we also use an alternative cutoff point of 33.3% in
defining the binary turnover variable to make sure our results are not driven by the choice of a
cutoff point, where the indicator variable equals one whenever at least one third of the top
management team is turned over, zero otherwise (turnover dummy II). Moreover, we verify the
robustness of our results by employing a third measure of turnover, management turnover ratio,
calculated as the percent of top management team turned over in a given year.
We exclude firms with total assets below $1 million and observations with missing
manager names or titles. We also drop years in which the size of the management team changes
by at least 50% to reduce the likelihood that our turnover variables are simply picking up any
potential reporting errors in the Amadeus database. After these screenings, we end up with
851,640 firm-year observations.
Listed firms make up 3% of this sample. Since there are significantly more private firms
than public firms in this sample, we use a matched sample of listed and unlisted firms in our
main analysis. However, we also perform robustness checks using the unmatched sample. The
procedure we follow in creating the matched sample is as follows. We first consider all listed
firms in 2008, as this year contains the largest number of firms for a given year in our sample.
We then match the listed firms to unlisted firms based on the country, industry (1-digit SIC) and
as close as possible on size measured by total assets.
9 We use a number of indexes that measure the development of countries’ institutions. We
use the antitakeover provision index from Nenova (2006) to measure the extent of anti-takeover
regulations across countries. To proxy for the scrutiny of public equity markets, we employ the
level of stock price synchronicity (R-squared) across countries from Jin and Meyers (2006).
1.2. Descriptive Statistics
Table 1 provides descriptive statistics for the three turnover measures we employ for
listed and matched unlisted firms. Panel A shows that on average unlisted firms experience less
turnover events than listed firms. Turnover dummy I has a mean of 0.09 and 0.11 for unlisted
and listed firms, respectively. These statistics suggest that on average every 11 (9) years, the
unlisted (listed) firms experience a turnover event where at least half of the top managers are
turned over. The difference is statistically significant at the one percent level. When we use the
alternative binary turnover variable indicating that at least one-third of the top management team
is turned over, the average duration increases to 5.5 (4.5) years. For comparison, Defond and
Hung (2004) report an average turnover ratio of 0.15 for listed firms around the world. We also
find averages similar to that of the second turnover dummy when we use the top management
turnover ratio as our measure of turnover event. Although listed firms appear to experience less
frequent turnover than unlisted firms based on the turnover ratio, panel B shows that all three
turnover measures are highly correlated with each other, with a minimum pairwise correlation of
77 percent.
Panel A also shows that listed firms are larger than unlisted firms in terms of total assets
($1,957 versus $1,064 million), have bigger management teams (3.45 versus 2.72) and somewhat
less profitable (0.06 versus 0.08). Panel C presents the number of observations for listed and
10 unlisted firms across countries. The largest numbers of observations in our sample are from the
U.K, France and Germany.
2. Turnover Analysis
2.1. Empirical Specification
To test the hypothesis that the sensitivity of top management turnover to poor
performance differs between listed and unlisted firms, we estimate a series of probit models that
take the following form:
,
,
∗
,
,
,
1
where Listed refers to listed firms and Management turnover is an indicator variable that takes
on the value of one whenever at least half of the top management team is turned over (turnover
dummy I). Our specification follows previous research such as Defond and Hung (2004) and
defines firm performance as the one-year lagged ratio of earnings before interest, taxes and
depreciation to total assets and includes a set of firm control variables, industry controls, and
year controls ( X ).We include country dummies to ensure we are measuring within-country
differences between listed and unlisted firms as well as controlling for unobserved country
effects. We also include industry and year dummies to control for industry wide factors and time
trends that may affect top manager turnover. We include an indicator variable that notes whether
the firm follows IFRS accounting standards to control for within country changes in financial
reporting since changing to IFRS can affect earnings measures (see, e.g. Ozkan, Singer, and You
(2013) and Daske, Hail, Leuz, and Verdi (2008)). Firm size, measured as the natural logarithm of
the book value of total assets in millions of U.S. dollars, is also added to control for the potential
11 effects of firm size on profitability and management turnover. . Finally, our regressions include
indicator variables for each year. We winsorize continuous variables at the one percent level and
correct the standard errors for possible serial correlation and heteroskedasticity by clustering at
the firm level. We take into account the non-linear nature of probit models in interpreting
interaction terms in our regressions (Norton, Wang and Ai (2004)). In a battery of robustness
test, we also consider alternative measures of management turnover and firm performance.
2.2 Sensitivity of Managerial Turnover to Poor Firm Performance
Results from the main specification that estimates the sensitivity of managerial turnover
to poor firm performance across listed and unlisted firms are reported Table 2, Model 1. The
coefficient of the interaction term, Listed*Lagged Earning Ratio, is negative and highly
statistically significant (-0.053, p=0.001) suggesting that listed firms are more likely to replace
poorly performing managers than unlisted firms. We also document that managerial turnover is
highly sensitive to firm performance in unlisted firms. The coefficient on the lagged earnings
ratio is -0.021 and it is statistically significant at the 1% level suggesting that unlisted firms are
more likely to replace top managers when performance is poor.
Among the control variables, firm size is positively related to managerial turnover. This
result is consistent with the earlier studies that examine cross-country samples of public firms
(Gibson (2003) and DeFond and Hung (2004)). Firms that follow IFRS display lower
management turnover, which is consistent with Sonali, Karpoff and Nahata (2012) who find that
CEO turnover is higher for firms that have lower reporting quality.
In Models 2 and 3 of Table 2, we examine robustness of our results to alternative
managerial turnover measures. We consider turnover dummy II (which equals one whenever at
12 least one third of the top management team is turned over and zero otherwise) in Model 2 and the
management turnover ratio in Model 3. We continue to find that listed firms display higher
sensitivity of managerial turnover to firm performance as the coefficients of the interaction terms
are -0.096 and -0.244, respectively, and both are statistically significant at the 1% level.
Taken together, the results in Table 2 suggest that top managers of public firms are less
likely to be entrenched than managers of private firms. Further, additional robustness checks
(untabulated) show that these results are not driven by financial firms, and remain when we
estimate our model using OLS.
2.3. The Market for Corporate Control and Scrutiny of Public Equity Markets
Our next set of results focus on the potential mechanisms that may drive the higher
sensitivity of management turnover in public firms. To this end, we exploit the cross-country
nature of our data and investigate how the propensity to replace poorly performing managers is
related to the development of countries’ institutions.
One potential mechanism that allows for replacement of poorly performing managers in
public firms is the market for corporate control. Manne (1965) and Jensen (1993), among others,
emphasize the importance of the market for corporate control in disciplining managers. For
example, Manne (1965) states that only the takeover market provides some assurance of
competitive efficiency among corporate managers and thereby affords strong protections to the
interests of vast numbers of small, non-controlling shareholders. Thus, the thrust of Berle and
Means’ famous phrase on the separation of ownership and control becomes less strong (p. 112113).
13 To test whether the market for corporate control is a mechanism that can explain our
findings, we use the antitakeover provision index from Nenova (2006) to measure the extent of
anti-takeover regulations across countries. More provisions that limit anti-takeover tactics
increase the likelihood of the replacement of poorly performing managers through the market for
corporate control. In Table 3, Panel A we run our base regression separately for countries with
strong and weak anti-takeover provisions. Model 1 present results for countries that limit
antitakeover tactics (the index values above median), while Model 2 presents results for
countries that do not limit antitakeover tactics (below the median). We find that listed firms are
more likely to replace poorly performing managers than unlisted firms only in countries that
limit antitakeover tactics thus facilitating the replacement of top managers.
Another potential governance mechanism for public companies is the scrutiny of public
equity markets. Information production and monitoring role of stock market in alleviating
agency problems in public corporations has been emphasized in a number of studies including
Fama (1980), Holmstrom and Tirole (1993), Dow and Gorton (1997), Subrahmanyam and
Titman (1999), and Edmans (2009). These studies suggest that different groups of market
participants, including not only current investors but also potential investors and financial
analysts, collect various information about public firms. Stock markets can aggregate this diverse
information across different investors and provide a useful signal that could be used to discipline
managers which could not have been obtained if the firm were privately financed.
To investigate whether the scrutiny of public equity markets is a mechanism that explains
our findings, we employ the level of stock price synchronicity (R-squared) across countries in
our tests. Morck, Yeung and Yu (2000) and Jin and Meyers (2006) show that an increase in
opaqueness leads to lower firm-specific risk for investors and higher R-squared. Panel B of
14 Table 3 shows that listed firms are more likely to replace poorly performing managers than
unlisted firms only in countries where stock markets produce more firm specific information
(i.e., have less stock market synchronicity). Taken together, our results on two important
mechanisms of public incorporation provide evidence that financial markets are not just a
sideshow but rather an important channel that limits managerial entrenchment and mitigates
agency problems in public corporations.
2.4 Alternative Explanations
2.4.1 Selection Effects
Our results demonstrate that public firms are more likely to replace poorly performing
managers than private firms which we interpret to suggest that listing status improves corporate
governance. A potential alternative explanation is that there is no advantage to being listed in
terms of the quality of corporate governance but rather well-governed firms choose to become
listed, perhaps for other reasons such as stock market liquidity. It is important to note that, even
if this alternative explanations is true, our results already highlight that going public doesn’t
cause the governance of these firms to degrade to the level that is below the governance level of
private firms as suggested by Berle and Means (1932).
Our first defense against this self-selection explanation was introduced by employing the
matched sample of listed and unlisted firms. Our sample selection procedure matches listed and
unlisted firms based on size and therefore in our tests we are picking unlisted firms that are more
like listed firms.6 Firm size is related to a number of important firm characteristics, including the
6
While one may consider matching on more firm characteristics, the number is tempered by statistical power
consideration.
15 quality of corporate governance.7 Nevertheless, to investigate further whether the self-selection
explains our results, we analyze firms that changed status from unlisted to listed or vice versa
within our sample period. If only well-governed firms opt to be listed and the change in status
doesn’t affect corporate governance, then there should be no difference in the sensitivity of top
management turnover to firm performance before and after the change in status.
The results from this test are presented in Model 1 of Table 4. We continue to find that
listed firm display higher sensitivity of management turnover to firm performance than unlisted
firms. This result suggests that the higher sensitivity of listed firms is likely to be driven by the
listing status rather than the selection effect.
We also investigate whether our results can be generalized to a broader set of private
firms. As described earlier, our matched sample includes the largest unlisted firms in a country.
To investigate whether our results are driven by these largest firms, we also consider the
unmatched sample of listed and unlisted firms, which includes predominantly small private
firms. Small private firms may differ from large private firm in the quality of corporate
governance. First, small private firms are likely to have more concentrated ownership than large
private firms and correspond closely to Jensen and Meckling’s (1976) 100% owner-manager
firms in which there is no separation of ownership and control (see Michael and Roberts (2011)
for detailed discussion). Second, unlike public listed firms, large private firms are not subject to
scrutiny of public equity markets.
Results from these regressions are presented in Model 2, Table 4. The findings suggest
that results in Table 2 are not driven by large unlisted firms. Listed firms are more likely to
replace poorly performing managers than unlisted firms, including small private firms.
7
Aggarwal, Erel, Stulz and Williamson (2009) show that a firm level governance index is highly correlated with
firm size (p. 3147).
16 2.4.2 Ownership Structure
We next examine the impact of differential ownership structure in public and private
firms in explaining our findings. First, we investigate whether high managerial ownership in
private firms may explain the relatively low likelihood of replacing poorly performing managers.
Denis, Denis and Sarin (1997) provide evidence that the probability of top executive turnover is
negatively related to managerial ownership. It is possible that managers of private firms have
relatively high stakes even in our matched sample that includes large private firms.8 Thus, in
Model 1 of Table 5, we consider a specification that excludes unlisted firms with high
managerial ownership such as family run private firms. We continue to find that listed firms are
more likely to replace poorly performing managers. This result further highlights potential
benefits associated with access to public equity markets.
As an additional test, we investigate how family ownership in public firms affects relative
managerial turnover. Findings in Villalonga and Amit (2006) suggest that agency conflicts are
quite severe in family-run public firms and, as a result, these firms might be less likely to replace
poorly performing managers. Indeed, we find the sensitivity of managerial turnover to firm
performance in family-run listed firms is similar to that in unlisted firms (Model 2 of Table 5).
This result suggests that the relative propensity to replace managers is related to the degree of
agency problems in public corporations.
2.4.3 Alternative firm performance measures
The third alternative explanation we examine is whether the lower sensitivity of top
management turnover to poor firm performance in private companies is driven by private firms’
8
We should note, however, that it is unclear whether higher managerial stakes in private firms necessarily lead to
entrenchment as other shareholders have higher stakes as well. 17 ability to consider longer performance windows in evaluating managers’ performance. Boot,
Gopalan and Thakor (2006, 2008) suggest a benefit of private ownership is that it enables the
manager to achieve the optimal level of decision-making discretion (or autonomy) through
private contacting with a few large investors, while such discretion is not visible in public firm
because of constantly changing investor base. Given a potentially higher level of managerial
discretion in private firms, it is probable that shareholders of private firms put more weight on
persistent poor performance in their decision to replace top managers. Therefore, we perform a
series of tests where performance measures are lagged up to three years.
In Model 1 of Table 6, we examine the sensitivity of the turnover to average firm
performance over the last three years which allows us to focus on a relatively long-term firm
performance rather than the performance over one year. The coefficient on the interaction term,
Listed*Firm Performance, remains negative and statistically significant suggesting that the
higher sensitivity of listed firms is unlikely to be driven by a relatively short-term firm
performance.
In Models 2 and 3, we further investigate whether unlisted firms display delayed reaction
in their decision to fire poorly performing managers. We consider lags two and three of the
earnings ratio. In both cases, the management turnover in unlisted firms is not sensitive to these
lagged performances. Both types of firms, listed and unlisted, are actually more likely to respond
to poor performance relatively quickly, within a year, although listed firms are more sensitive not
only to a firm performance over the previous year but also to a firm performance that is lagged
two years. Overall, the results in Table 2 are unlikely to be due to the delayed response by
unlisted firms.
18 In untabulated regressions, we also consider earnings before interest and taxes, return on
assets, sales growth and cash flow from operations as measures of firm performance. We run
specifications similar to those in Table 2 and find no evidence that public firms display lower
sensitivity of top management turnover to firm performance than private firms.
3. Conclusion
Diffused ownership, a salient feature of modern public corporations, is often argued to lead to
substantial agency conflicts in public firms as it makes it more difficult to replace poorly
performing managers. Private firms, on the other hand, typically have concentrated ownership
and control, but are less likely to be subject to governance mechanisms associated with public
equity markets. In this paper, we provide a comparison of the severity of agency problems
between public and private firms by employing a primary outcome of corporate governance, the
propensity to dismiss poorly performing managers. Our sample includes European countries,
which subject their private firms to similar reporting requirements as public firms thereby
allowing us to benchmark public firms with a comprehensive group of private firms.
We find that public firms are significantly more likely to replace poorly performing top
managers than private firms in our cross-country sample, pointing to a stronger governance
environment in public firms. This result is surprising in the framework of Berle and Means
(1932). Our results also show that this increased sensitivity of management turnover to firm
performance at public firms is related to the information production and monitoring role of stock
markets and to the market for corporate control. These results indicate that public firms have
better checks and balances in place than private firms to mitigate agency problems, in part
because of the important governance function of public equity markets.
19 References
Aggarwal, R., I. Erel, M. Ferreira, and P. Matos. 2011. Does Governance Travel Around the
World? Evidence from Institutional Investors, Journal of Financial Economics 100, 154-181.
Aggarwal, R., I. Erel, R. Stulz, and R. Williamson, 2009, Differences in Governance Practices
between U.S. and Foreign Firms: Measurement, Causes, and Consequences, Review of Financial
Studies 22, 3131- 3169.
Berle, A., and G. Means. 1932. The Modern Corporation and Private Property. New York, NY:
Macmillan.
Boot, A., R. Gopalan, and A. Thakor, 2006, The Entrepreneur’s Choice between Private and
Public Ownership, Journal of Finance 61, 80-836.
Boot, A., R. Gopalan, and A. Thakor, 2008, Market Liquidity, Investor Participation, and
Managerial Autonomy: Why Do Firms Go Private, Journal of Finance 63, 2013-2059.
Coles, H., M. Lemmon, and L. Naveen. 2003. A Comparison of Profitability and CEO Turnover
Sensitivity in Large Public and Private Firms. Working paper.
Daske, H., L. Hail, C. Leuz, and R. Verdi. 2008. Mandatory IFRS Reporting around the World:
Early Evidence on the Economic Consequences. Journal of Accounting Research 46, 1085-1142.
Defond, M., and M. Hung. 2004. Investor Protection and Corporate Governance: Evidence from
Worldwide CEO Turnover. Journal of Accounting Research 42, 269-312.
Denis, D., D. Denis, and A. Sarin, 1997. Ownership Structure and Top Executive Turnover.
Journal of Financial Economics 45, 193-221.
Dow, J. and G. Gorton. 1997. Stock Market Efficiency and Economic Efficiency: Is There a
Connection? Journal of Finance 52, 1087-1129.
Edgerton, J. 2012. Agency Problems in Public Firms: Evidence from Corporate Jets in Leverage
Buyouts, Journal of Finance 67, 2187–2213.
Edmans, A. 2009. Blockholder Trading, Market Efficiency, and Managerial Myopia. Journal of
Finance 64, 2481–2513.
Edmans, A., I. Goldstein, and W. Jiang. 2012. The Real Effects of Financial Markets: The
Impact of Prices on Takeovers. Journal of Finance, forthcoming.
Fama, E. 1980. Agency Problems and the Theory of the Firm. Journal of Political Economy 88,
288–07.
Giannetti, M. 2003. Do Better Institutions Mitigate Agency Problems? Evidence from Corporate
Finance Choices, Journal of Financial and Quantitative Analysis 38, 185-212.
20 Gibson, M. 2003. Is Corporate Governance Ineffective in Emerging Markets? Journal of
Financial and Quantitative Analysis 38, 231–250.
Gupta, N., 2005. Partial Privatization and Firm Performance, Journal of Finance 60, 987-1015.
Hazarika, S., J. Karpoff, and R. Nahata. 2012. Internal Corporate Governance, CEO Turnover,
and Earnings Management. Journal of Financial Economics 104, 44-69.
Hermalin, B. and M. Weisbach. 2003. Boards of Directors as an Endogenously Determined
Institution: A Survey of the Economic Literature. Economic Policy Review 9, 7-26.
Holmstrom, B., and J. Tirole. 1993. Market Liquidity and Performance Monitoring. Journal of
Political Economy 101, 678-709.
Jensen, M., 1993. The Modern Industrial Revolution, Exit, and the Failure of Internal Control
Systems. Journal of Finance 48, 831-880.
Jensen, M. and W. Meckling. 1976. The Theory of the Firm: Managerial Behavior, Agency Cost,
and Ownership Structure. Journal of Financial Economics 3, 305-360.
Jin, L., and S. Myers. 2006. R2 around the World: New Theory and New Tests, Journal of
Financial Economics 79, 257-292.
Lel, U., and D. Miller. 2008. International Cross-Listing, Firm Performance, and Top
Management Turnover: A Test of the Bonding Hypothesis. Journal of Finance 63, 1897-1937.
Manne, H. 1965. Mergers and the Market for Corporate Control, Journal of Political Economy
73, 110-120.
Michaely, R. and M. Roberts. 2012. Corporate Dividend Policies: Lessons from Private
Firms, Review of Financial Studies 25, 711-746.
Mikkelson, W. and M. Partch. 1997. The Decline of Takeovers and Disciplinary Managerial
Turnover. Journal of Financial Economics 44, 205-239.
Morck, R., B. Yeung, and W. Yu. 2000. The Information Content of Stock Markets: Why Do
Emerging Markets Have Synchronous Stock Price Movements? Journal of Financial Economics,
215-260.
Mortal, S. and N. Reisel, 2013. Capital Allocation by Public and Private Firms. Journal of
Financial and Quantitative Analysis 48, 77-103.
Nenova, T. 2006. Takeover Laws and Financial Development. World Bank Policy Research
Working Paper 4029.
Norton, E., H. Wang, and C. Ai. 2004. Computing Interaction Effects and Standard Errors in
Logit and Probit Models. Stata Journal 2, 103 – 116.
21 Ozkan, N., Z. Singer, and H. You, 2013. Mandatory IFRS Adoption and the Contractual
Usefulness of Accounting Information in Executive Compensation. Journal of Accounting
Research, forthcoming.
Sheen, A. 2009. Do Public and Private Firms Behave Differently? An Examination of
Investment in the Chemical Industry. Working paper.
Shleifer, A. and R. Vishny. 1989. Management Entrenchment: The Case of Manager-Specific
Investment. Journal of Financial Economics 25, 123–140.
Shleifer, A. and R. Vishny. 1997. A Survey of Corporate Governance, Journal of Finance 52,
737–783.
Subrahmanyam, A., and S. Titman. 1999. The Going-Public Decision and the Development of
Financial Markets. Journal of Finance 54, 1045-1082.
Villalonga, B., and R. Amit. 2006. How Do Family Ownership, Control and Management Affect
Firm Value? Journal of Financial Economics 80, 385-417.
Volpin, P. 2002. Governance with Poor Investment Protection: Evidence from Top Executive
Turnover in Italy, Journal of Financial Economics 64, 61–91.
Table 1. Descriptive Statistics
22 This table presents descriptive statistics for the main variables used in the analysis. The sample includes
public and private firms from Western European countries. Details of the sample selection procedure are
provided in the text. Listed is one for public firms listed on major exchanges, and zero otherwise. Unlisted
is one for private firms or firms with the public legal status that are not listed on major exchanges, and
zero otherwise. Top management turnover dummy I equals one for firms when at least 50 percent of the
top management team is turned over in a given year. Top management turnover dummy II equals one for
firms when at least 1/3 of the top management team is turned over in a given year. Top management
turnover ratio is the percent of top management team that is turned over in a given year. Lagged Earnings
Ratio is the one-year lagged ratio of earnings before interest, taxes and depreciation to total assets. Total
Assets are measured in millions of $US. IFRS dummy equals one for firms that follow the IFRS
accounting standards, and zero otherwise. Asterisks ***, **, and * indicate significance at the 1%, 5%,
and 10% level, respectively.
Panel A. Means
Top management
turnover dummy I
Top management
turnover dummy II
Top management
turnover ratio
Total number of top
managers
Lagged Earnings
Ratio
Total Assets
Unlisted
N
Listed
N
Diff. in
means
0.09
20,442
0.11
18,754
-0.02***
0.18
21,083
0.22
19,973
-0.04***
0.20
24,509
0.16
20,608
0.04***
2.72
24,509
3.45
20,608
-0.73***
0.08
22,008
0.06
19,509
0.02***
1,064.52
24,509
1,957.56
20,608
-893.04***
Panel B. Correlations
1. Top management turnover dummy I
2. Top management turnover dummy II
3. Top management turnover ratio
1
1.00
0.77
0.84
23 2
3
1.00
0.86
1.00
Table 1 (continued)
Panel C. Number of observations across countries
Austria
Belgium
Denmark
Finland
France
Germany
Greece
Iceland
Ireland
Italy
Norway
Portugal
Spain
Sweden
United Kingdom
Unlisted
Listed
34
271
538
784
5,263
2,907
2,040
24
258
30
910
189
1,662
2,049
7,559
30
151
478
896
3,610
3,182
2,186
16
278
13
637
137
845
1,533
6,616
24 Table 2. Top Management Turnover and Listing Status: Main Specification and Alternative
Turnover Ratios
This table presents the Probit and Tobit estimates of the relationship between the top management
turnover measures and firm performance. The average marginal effects are reported. The interaction
terms in Probit regressions are estimated using the methodology of Norton, Wang, and Ai (2004); the
average effects are reported. The sample includes public and private firms from Western European
countries. Details of the sample selection procedure are provided in the text. The dependent variable is top
management turnover dummy I in column (1), top management turnover dummy II in column (2), and top
management turnover ratio in column (3). IFRS dummy equals one for firms that follow the IFRS
accounting standards, and zero otherwise. Other variable are described in Table 1. The z-statistics appear
in parentheses below parameter estimates. Robust standard errors are estimated using Rogers method of
clustering by firm. Asterisks ***, **, and * indicate significance at the 1%, 5%, and 10% level,
respectively.
Variable
Listed
Listed * Lagged Earnings Ratio
Lagged Earnings Ratio
Log Assets
IFRS dummy
Industry Dummies
Year Dummies
Country Dummies
N
Log Pseudolikelihood
(1)
(2)
(3)
0.017***
(4.62)
-0.053***
(-3.85)
-0.021***
(-2.79)
0.005***
(5.83)
-0.016***
(-3.15)
0.045***
(9.16)
-0.096***
(-4.99)
-0.017*
(-1.62)
0.012***
(10.51)
-0.015**
(-2.34)
-0.066***
(-6.07)
-0.244***
(-6.45)
-0.022
(-1.10)
0.040***
(14.88)
0.029**
(2.20)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
36,146
-11,180.434
37,829
-18,268.003
41,517
-31,580.882
25 Table 3. Top Management Turnover and Listing Status: Cross-country Analysis
This table presents the Probit estimates of the relationship between the top management turnover measure
and firm performance. The average marginal effects are reported. The interaction terms in Probit
regressions are estimated using the methodology of Norton, Wang, and Ai (2004); the average effects are
reported. The sample includes public and private firms from Western European countries. Details of the
sample selection procedure are provided in the text. The dependent variable is top management turnover
dummy I. The antitakeover provisions index is from Nenova (2006). The stock market scrutiny is
measured using average R-squared from Jin and Myers (2006). In all panels, specification (1) reports
results for the index above median, while specification (2) reports result for the index below median.
IFRS dummy equals one for firms that follow the IFRS accounting standards, and zero otherwise. Other
variable are described in Table 1. Robust standard errors are estimated using Rogers method of clustering
by firm. Asterisks ***, **, and * indicate significance at the 1%, 5%, and 10% level, respectively.
Panel A. Anti-takeover provisions
Variable
Listed
Listed * Lagged Earnings Ratio
Lagged Earnings Ratio
Log Assets
IFRS dummy
Industry Dummies
Year Dummies
Country Dummies
N
Log Pseudolikelihood
The anti-takeover index
above median
(1)
The anti-takeover index
below median
(2)
0.013***
(3.26)
-0.042***
(-2.89)
-0.024***
(-2.61)
0.005***
(5.06)
-0.013***
(-2.46)
0.026**
(2.14)
0.038
(0.39)
-0.011
(-1.45)
0.001
(0.27)
-0.005
(-0.26)
Yes
Yes
Yes
Yes
Yes
Yes
31536
-9709.725
2137
-431.438
26 Table 3 (continued)
Panel B. Stock market scrutiny
Variable
Listed
Listed * Lagged Earnings Ratio
Lagged Earnings Ratio
Log Assets
IFRS dummy
Industry Dummies
Year Dummies
Country Dummies
N
Log Pseudolikelihood
The stock market scrutiny
above median
(1)
The stock market scrutiny
below median
(2)
0.038***
(5.13)
0.005
(0.11)
-0.031
(-1.18)
0.002
(1.30)
-0.001
(-1.39)
0.013***
(2.86)
-0.056***
(-3.82)
-0.020***
(-2.49)
0.005***
(4.73)
-0.026***
(-3.68)
Yes
Yes
Yes
Yes
Yes
Yes
7,181
-1,810.956
25,104
-7,884.85
27 Table 4. Top Management Turnover and Listing Status: Alternative samples
This table presents the Probit estimates of the relationship between the top management turnover measure
and firm performance. The average marginal effects are reported. The interaction terms in Probit
regressions are estimated using the methodology of Norton, Wang, and Ai (2004); the average effects are
reported. The sample includes public and private firms from Western European countries. Details of the
sample selection procedure are provided in the text. The dependent variable is top management turnover
dummy I. Specification (1) reports results for the sub-sample of firm that changed status from listed to
unlisted or vice versa. Specification (2) reports results for unmatched sample. IFRS dummy equals one
for firms that follow the IFRS accounting standards, and zero otherwise. Other variable are described in
Table 1. The z-statistics appear in parentheses below parameter estimates. Robust standard errors are
estimated using Rogers method of clustering by firm. Asterisks ***, **, and * indicate significance at the
1%, 5%, and 10% level, respectively.
Variable
Listed
Listed * Lagged Earnings Ratio
Lagged Earnings Ratio
Log Assets
IFRS dummy
Industry Dummies
Year Dummies
Country Dummies
N
Log Pseudolikelihood
(1)
(2)
0.046***
(8.44)
-0.029**
(-1.98)
-0.004
(-0.34)
0.005***
(4.18)
-0.011*
(-1.77)
0.000
(0.20)
-0.060***
(-4.44)
-0.006**
(-2.08)
0.015***
(55.95)
0.006***
(2.77)
Yes
Yes
Yes
Yes
Yes
Yes
18,878
-5,634.73
645,818
-181,868.54
28 Table 5. Top Management Turnover and Listing Status: Ownership
This table presents the Probit estimates of the relationship between the top management turnover measure
and firm performance. The average marginal effects are reported. The interaction terms in Probit
regressions are estimated using the methodology of Norton, Wang, and Ai (2004); the average effects are
reported. The sample includes public and private firms from Western European countries. Details of the
sample selection procedure are provided in the text. The dependent variable is top management turnover
dummy I. In specification (1), the sample of unlisted firms excludes family-run firms. In specification (2),
the sample of listed firms is limited to family-run firms. IFRS dummy equals one for firms that follow the
IFRS accounting standards, and zero otherwise. Other variable are described in Table 1. The z-statistics
appear in parentheses below parameter estimates. Robust standard errors are estimated using Rogers
method of clustering by firm. Asterisks ***, **, and * indicate significance at the 1%, 5%, and 10% level,
respectively.
Variable
Listed
Listed * Lagged Earnings Ratio
Lagged Earnings Ratio
Log Assets
IFRS dummy
Industry Dummies
Year Dummies
Country Dummies
N
Log Pseudolikelihood
(1)
(2)
0.012***
(3.04)
-0.052***
(-3.79)
-0.021***
(-2.69)
0.004***
(4.85)
-0.018***
(-3.36)
-0.002
(-0.37)
-0.024
(-0.86)
-0.016**
(-2.33)
0.009***
(8.31)
-0.003
(-0.41)
Yes
Yes
Yes
Yes
Yes
Yes
34,098
-10,817.833
21,889
-6,416.1502
29 Table 6. Top Management Turnover and Listing Status: Alternative Firm Performance Measures
This table presents the Probit estimates of the relationship between the top management turnover measure
and firm performance. The average marginal effects are reported. The interaction terms in Probit
regressions are estimated using the methodology of Norton, Wang, and Ai (2004); the average effects are
reported. The sample includes public and private firms from Western European countries. Details of the
sample selection procedure are provided in the text. The dependent variable is top management turnover
dummy I. Average Earnings Ratio is the mean Earnings Ratio over the last three years. IFRS dummy
equals one for firms that follow the IFRS accounting standards, and zero otherwise. Other variable are
described in Table 1. The z-statistics appear in parentheses below parameter estimates. Robust standard
errors are estimated using Rogers method of clustering by firm. Asterisks ***, **, and * indicate
significance at the 1%, 5%, and 10% level, respectively.
Variable
Listed
Listed * Average Earnings Ratio
Average Earnings Ratio
(1)
(2)
(3)
0.016***
(4.34)
-0.033*
(-1.83)
-0.024***
(-3.43)
0.015***
(3.66)
0.014***
(3.01)
-0.031**
(-1.99)
-0.009
(-1.04)
Listed * Lag Two Earnings Ratio
Lag Two Earnings Ratio
Listed * Lag Three Earnings Ratio
0.005***
(5.52)
-0.017***
(-3.35)
0.006***
(5.97)
-0.016***
(-3.00)
-0.200
(-1.51)
-0.006
(-0.79)
0.005***
(4.67)
-0.012**
(-2.19)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
36,761
-11,376.017
29,339
-8,785.046
24,001
7,320.173
Lag Three Earnings Ratio
Log Assets
IFRS dummy
Industry Dummies
Year Dummies
Country Dummies
N
Log Pseudolikelihood
30