5
1. Introduction
CEO removals are considered crucial events in order to understand those governance
mechanisms which discipline managers, avoiding deviation from value-maximizing corporate
policies (Jensen and Warner (1988)). If all economic systems face the same principal-agent
problem, CEO changes should also represent important disciplinary forces for different
countries.
This thesis focuses on CEO changes, providing an analysis of differences in governance
mechanisms, company performance and corporate restructuring activities in the U.K. and
Germany.
Early studies on managerial turnover by Coughlan and Schmidt (1985), Warner, Watts and
Wruck (1988), Weisbach (1988) and Jensen and Murphy (1990) suggest turnover being
negatively related to past performance. More recently, Kaplan (1994b), Renneboog (2000),
Franks, Mayer, and Renneboog (2001), Volpin (2002) and Maury (2006) confirm this
evidence. However, most of the reference literature on CEO turnover has focused on
single-country studies with the only exception of Kaplan (1994a)
1
.
The main contribution of this thesis is to search for cross-country evidence of the causes
and effects of CEO changes. In detail, I examine whether negative past performance increases
the likelihood of CEO turnover in both countries. I investigate which governance mechanisms
influence managerial turnover under different governance regimes, and if their influence
increases for underperforming firms. Moreover, I search for cross-country evidence in terms
of post-turnover performance reversals and corporate restructuring.
I have taken the U.K. and Germany since they represent two opposite legal and economic
systems, displaying vast differences in matters such as investor protection, financial
1
Kaplan (1994a) analyses corporations from the U.S. and Japan and finds that negative past performance
increases the likelihood of CEO turnover in a similar way in both countries.
6
development and employment legislation
2
. The U.K. and Germany also show differences in
their corporate governance regimes, indeed the U.K. displays one-tier boards while Germany
features two-tier boards. Moreover, German firms are subject to a codetermination system,
whereby some of the members of the supervisory board are appointed by employees
3
.
I look at four potential corporate governance mechanisms to keep CEO agency conflicts in
check: Large shareholders, board structure, takeover pressure and capital structure.
The U.K. and Germany show differences in all their roles.
Considering ownership structure, the average U.K. firm is widely held while the average
German corporation has a controlling blockholder and a higher ownership concentration
(Faccio and Lang (2002) and Franks, Mayer, Volpin and Wagner (2009)). Looking at board
structure, I focus on the role of outside directors and board size. Outside directors are seen as
a way to increase turnover-sensitivity to past performance in the U.S. (Weisbach (1988)), but
evidence for the U.K. is mixed (Dahya, McConnell, and Travlos (2002) and Franks, Mayer,
and Renneboog (2001)). For Germany, a comparative study of the independence of German
outside directors with their U.S. and U.K. counterparts has been made by Roe (1993) and
Fauver and Fuerst (2006). Considering board size, Dahya, McConnell and Travlos (2002)
show that smaller boards are better CEO monitors in the U.K. For Germany, board size is
regulated by law with a minimum of 3 and a maximum of 21 members, with the actual
number depending on company size and industry sector
4
. Regarding takeover pressure, the
market for corporate control conclusively plays a monitoring role in the U.K. while in
2
Considering investor protection, research by La Porta, Lopez-de-Silanes, Shleifer and Vishny (1998), by La
Porta, Lopez-de-Silanes and Shleifer (2006) and by Djankov, La Porta, Lopez-de-Silanes and Shleifer (2008)
suggest that the U.K. system provides a stronger investor protection than the German one. The U.K. is also
characterized by a more developed financial system than Germany. Indeed, in 2002 London was the fourth
largest stock market in the world in terms of market value of domestic equity and third largest in terms of
domestic turnover, while the Deutsche Borse ranked sixth both for market value of domestic equity and domestic
turnover (Buckle and Thompson (2004)). If we consider employment protection, Botero, Djankov, La Porta, and
Lopez-de-Silanes (2004) suggest workforce protection is stronger in Germany than in the U.K.
3
On this point Roe (1993) advocates that excessive codetermination may reduce German supervisory board
effectiveness while Fauver and Fuerst (2006) argue that the German codetermination system leads to better
board decisions.
4
See Paragraph 95 of Aktiengesetz.
7
Germany, at least at the beginning of my sample period (1995), the takeover market is widely
regarded as inactive (OECD (1998) and Franks, Mayer, Volpin, and Wagner (2009)). Finally,
German and U.K. corporations exhibit different financing sources, with bank financing
playing a larger role in German firms, often in the form of relationship banking (Franks and
Mayer (2001)).
Do these difference matter? I find that, although each country-regime relies on different
sets of governance mechanisms, their combined effect is remarkably similar. Results suggest
that managers are held accountable for poor performance even if they operate under very
different governance systems.
The analysis of turnover determinants suggests CEO turnover in the U.K. is influenced by
board independence, board size, takeover pressure and debt. For Germany, I find relevant
governance mechanisms to be large shareholders presence and debt. Leverage therefore plays
an important role in both countries as it increases the likelihood of CEO removal, but I find
this effect being larger for Germany. This result is in line with creditors playing an additional
monitoring role in those firms which have a sensible portion of bank financing.
Findings on the disciplinary impact of governance mechanisms indicate that debt alone
plays a disciplinary role, and only in the case of German firms at that, thus increasing the
odds of CEO turnover for underperforming firms. Evidence on the other factors suggests they
operate in an unfocused way in both countries.
Regarding post-turnover performance changes, I find that in both countries CEO removal
in poorly performing firms leads to performance improvements, while the opposite holds
when well-performing CEOs a are replaced. These results are robust to adjusting for industry
and mean reversion, as captured by control group adjustment performed following the Barber
and Lyon (1996) method. In the case of the U.K., this evidence replicates results obtained for
the U.S. by Denis and Denis (1995) and Huson, Malatesta and Parrino (2004). Regarding
8
Germany, my results are novel and complement those of Kaplan (1994b) who investigates the
relation between managerial turnover and past performance but does not consider
post-turnover performance.
If we consider post-turnover restructuring activities in the U.K. and Germany, companies
undergoing a disciplinary turnover (i.e. a CEO removal in an underperforming firm) and a
non-disciplinary turnover, exhibit a significant downsizing in assets in the three-year period
following the removal of the CEO. Such a reduction is larger for disciplinary turnovers than
for non-disciplinary turnovers and is of similar size in both countries. Remarkably, this
pattern holds also when workforce restructuring is considered. This finding is particularly
interesting given that German workers enjoy a higher level of protection than their U.K.
counterparts and can affect board decisions via the codetermination system.
My analysis relies on previous research by Dimopoulos and Wagner (2009) who compare
cause and effect in CEO changes for the U.K. and Germany. In detail, my contribution to
work by Dimopoulos and Wagner (2009) is increasing the German sample from 150 to 300
companies and updating accounting and employment data adopting the newly introduced
version of COMPUSTAT Global. I also implement in a different way industry and control
group adjustment procedures exploited by Dimopoulos and Wagner (2009). My research
relies on a specifically-constructed data panel obtained by matching information from a
number of data sources. Among these are COMPUSTAT, Hoppenstedt Aktienfuehrer,
Manifest U.K., Thomson One Banker, Factiva and Capital IQ. The overall sample comprises
6,800 firm-year observations for the U.K. and 2,700 for Germany. Regarding Germany, my
sample stands as the broadest firm sample used so far in research on CEO turnover.
The remainder of the thesis is structured as follows. Section 2 discusses hypotheses and
methodology along with reference literature. Section 3 discusses the data. Section 4 shows
and comments the empirical findings, while Section 5 concludes the thesis.
9
2. Hypotheses and methodology
2.1 Hypotheses
In this section I report theoretical hypotheses along with a discussion of the reference
literature. The hypotheses concern the relation between managerial turnover and company
past performance. In particular, the way in which different corporate governance mechanisms
can affect turnovers and the relation between turnover events and company post-turnover
performance.
Hypotheses testing results are reported in section 4.
2.1.1 Managerial turnover and company past performance
(H1) A negative relation exists between company past performance and the likelihood of
consequent CEO turnover in both the U.K. and Germany.
Theory suggests that bad performance causes a reputation loss for the CEO and increases
the probability that the board of directors replaces him (Hermalin and Weisbach (1998)).
The relation between company past performance and turnover events has also attracted a
notable portion of prior research on executive turnover.
Coughlan and Schmidt (1985), Warner, Watts and, Wruck (1988) and Jensen and Murphy
(1990) prove turnover to be inversely related to firm performance
5
.
5 Coughlan and Schmidt (1985) analyse a panel of 249 U.S. listed corporations between 1978 and 1982
finding a significant negative relation between abnormal corporate stock returns and the probability of
subsequent CEO turnover. Warner, Watts and, Wruck (1988) and Jensen and Murphy (1990) also confirm this
evidence. The first study finds CEO, President, and Chairman turnover is inversely related to stock performance
for a panel of 269 random firms listed on the NYSE and on the AMEX, which they track over a period from
1962 to 1978. Moreover, they suggest that stock performance relative to the market is a better predictor of
turnover than stock return only. Likewise, Jensen and Murphy (1990) find CEO turnover is negatively related to
past performance as measured by realized shareholders’ return in excess of NYSE return. Their results rely on a
panel of 1,049 U.S. listed corporations from 1974 to 1986.
10
CEO turnover is not negatively related to stock performance only, but it also shows an
inverse relation to accounting performance measures. Weisbach (1988) finds CEO turnover to
be negatively related to both stock returns and changes in EBIT for firms whose board
consists mainly of outside directors
6
. The decision to consider also an accounting
performance measure rather than stock returns only, relies on the fact that these two types of
performance metrics embed different information. Indeed, while stock prices reflect the
present discounted value of the expected future cash flows of the company, earnings data
represent short term profits and should therefore display a stronger connection to turnover
events.
More recently, Fee and Hadlock (2004) find evidence that the forced turnover rate for
non-CEO top executives is as high as that of CEOs even though their sensitivity to past
performance is lower, while Kaplan and Minton (2006) document that CEO turnover has
increased its sensitivity to past performance in recent years.
Evidence pointing to a negative relation between firm performance and CEO turnover is
also confirmed by an international body of single-country studies
7
.
Therefore, my rationale is to prove that a negative relation exists between company past
performance and CEO turnover in both the U.K. and Germany.
6 Weisbach (1988) findings rely on an analysis of 495 publicly-held corporations listed on the NYSE between
1977 and 1980.
7 Kaplan (1994b) confirms turnover in German firms increasing significantly with poor past performance.
Still, his analysis relies on a relatively small sample of 42 large firms over a period going from 1981 to 1989.
The connection between turnover and poor performance holds also for Belgian companies as proven by
Renneboog (2000), who analyses all firms listed on the Brussels Stock Exchange from 1989 to 1994. A similar
research for U.K. corporations is by Franks, Mayer, and Renneboog (2001), who analyse 250 random firms
quoted on the London Stock Exchange from 1988 to 1993. Finally, Volpin (2002) and Maury (2006) document a
negative correlation between CEO turnover and firm performance for companies in Italy and in Finland
respectively.
Ricevi informazioni sui nostri servizi, sulle offerte e non perdere news e consigli su università e lavoro.
