Manager turnover and product market competition:
evidence from UK family firms.
Su Wang*
London School of Economics
This draft: September 2016
* Department
of Finance, Houghton Street, London WC2A 2AE, UK, E-mail: [email protected].
Manager turnover and product market competition:
evidence from UK family firms.
Abstract
This paper investigates the impact of product market competition from foreign markets on
top managerial position changes in family firms. We estimate the causal effect by using import
penetration to measure foreign competition and implementing import weighted exchange rates
index as instrumental variable. Using senior executive manager data of UK family firms in
manufacturing sector, we find that intense foreign competition causes significant increase in the
frequency of family manager resignations without replacement and the number of unrelated
replacements after unrelated manager departures. Family firms with more unrelated managers
replacing family or unrelated managers are more likely to be acquired in the future. Furthermore,
family firms without founders serving as current managers are more silent in face of foreign
competition. Overall, these results suggest that we can not reject the joint hypothesis: 1) family
transitions provide private benefits of control, and 2) product market competition changes the
mapping between the profits and the managerial effort/skill.
1
Introduction
Family firms are becoming the important and major players in the economy around the world,
especially in Europe and Asia (Porta, Lopez-de Silanes, and Shleifer (1999); Claessens, Djankov,
and Lang (2000); Faccio and Lang (2002)). In UK, family business made a £346 billion valueadded contribution to GDP in 2011, providing 40% of total private sector employment (UK
family business sector report Nov 2011 Final). Despite the prevalence of family firms and
the ambiguous performance of family firms compared with non-family firms in light of recent
researches, the corporate governance inside family firms remains understudied.
The senior executive manager succession decision is a major and critical challenge in determining the frim growth in the long run and survival in difficult times. It is a particular topical
issue for the family companies. The choice of either appointing a connected family member or
selecting an unrelated person to the key managerial positions can affect the corporate policies
such as investment decisions and hence influnce the firm performances (Bennedsen, Nielsen,
Perez-Gonzalez, and Wolfenzon (2007); Tsoutsoura (2015)). This paper empirically investigates
how family firms balance the benefits and costs of the resignation and the promotion of family
member to key corporate management positions. Specifically, we show that product market
competition from foreign markets is one force significantly affect the succession plans and the
structure of top executive management team in family firms.
In theory, product market competition has long being viewed as a vehicle to mitigate managerial slack in corporations (Hart (1983); Schmidt (1997)). Faced with intense product market
competition, unrelated manager has more incentive to exert effort and reduces the monitoring
cost. Unrelated managers with better skills selected from a larger talent pool are more capable
than the family managers, thus favouring the choice of the unrelated manager. However, family
managers have less myopic investment strategies compared to unrelated managers. Furthermore,
the good reputation carried on by the founder or the previous family managers, the business
relationship built up, and the private benefits of family control add to advantages of passing
the firm onto family hands. Altogether, theory suggests an ambiguous prediction of manager
turnovers with increasing competition.
2
Recognized by previous reserches, the measurement of product competition such as concentration ratio, Herfindahl indices and price-cost margins are difficult to compute and interpret.
Those measures also raise the concern of endogeneity. To avoid the flaws of those measurements,
we compute the import penetration at industry level to measure the product market competition from foreign markets. United Kingdom is the fifth largest import country in the world,
especially in manufactured goods. This makes foreign competition a crucial channel through
which the product market competition changes the reward to exerting effort or hiring more
talented managers and the relative advantages of hiring a family member varies accordingly.
Furthermore, foreign competition is an important force that can have substantial impact on
firm level outcomes. It has been showed in empirical literature that competition from foreign
markets spurs innovation, improves productivity, changes the ownership structure and improves
the management quality (Bloom, Draca, and Van Reenen (2016); Bena and Xu (2016); Bloom,
Propper, Seiler, and Van Reenen (2015); Cuñat and Guadalupe (2009)). To address the potential
endogenity concerns because of either the measurement error or the family firms’ anticipation
of import penetration movements, we follow Bertrand (2004); Cuñat and Guadalupe (2009) and
instrument the import penetration with the import weighted average exchange rates at industry
level.
Our study draws on historical managerial data from Orbis database provided by the Bureau
van Dijk (BvD). We identify the family firms by matching the family names of senior executive
managers. Using the information of appointment date and the resignation date of senior executive managers, we are able to track the change of senior executive managerial positions for
each firm over time. In addition, our data provides the financial information for most private
family firms in sample, allowing us to control for firm characteristics in the subsequent analysis.
Trade data and domestic production data is collected from the Eurostat database. We focus on
the sample of UK manufacturing family firms because of the availability of domestic production
data. We use data from 2004 to 2014, which covers the Pounds appreciation period around 2006
and 2012, generating the time varying exchange rates index at the industry level.
3
Our findings show that jointly with firm size and age, product market competition from the
foreign markets places an important role in senior executive managerial position changes. Foreign
competition substantially increases the frequency of senior executive managers resignations (with
or without replacement), and specifically, the number of family executives departures. Our
results also show that family firms fire more family executives without replacement and they
change the current unrelated managers with new unrelated ones. In the subsample analysis
over founder family firms and nonfounder family firms, we show that the main results hold for
the founder family firm sample. However, family firms without the family founder as current
manager are more quite when faced with high level of foreign competition. To explore the
subsequent outcomes after managerial position changes, we test the likelihood of merge and
acquisition event and find that the number of unrelated replacement after family or unrelated
departures are positively associated with the probability of being acquired in the following two
years.
This paper adds to the literature in two strands. Firstly, instead of looking at the family
control on firm performance (Anderson and Reeb (2003); Villalonga and Amit (2006a); Ellul,
Pagano, and Panunzi (2010); Bennedsen, Nielsen, Perez-Gonzalez, and Wolfenzon (2007), etc.),
we are attempting to answer the question of what factors drive the managerial position changes.
This paper complements the existing researches in the way that we find that product market
comeptition is one of the important economic factor that has significant impact on the succession decisions. Secondly, different from literature that focuses on the relationship between
competition and corporate governance of public firms (Guadalupe and Pérez-González (2010);
Kadyrzhanova and Rhodes-Kropf (2011); Giroud and Mueller (2010, 2011)), we extend the investigation of the product market competition and its impact on management succession changes
based on a sample of private UK family firms.
The rest of the paper is as follows: section 2 presents the motivation and the related
literature; section 3 describes the data; section 4 outlines our empirical strategy; section 5 shows
the results of the paper and section 6 concludes.
4
2
Motivation
From the Santander bank succession gap to the Sumsung group“soft succession”, family control has drawn much attention from the public. Family control brings several advantages to
family firms according to theoretical models. Firstly, family management mitigates the classical
principal agent problem (Shleifer and Vishny (1986); Bolton, Thadden, et al. (1998)) between
the owner and the manager because family owners usually manage the firm themselves. Secondly, family companies prompt the long-term firm specific investments to overcome the typical
asset substitution problems. Lastly, they carry on the reputation and the culture which help
to run the firm more smoothly over difficult times. However, family managers may distort the
objective of maximizing the firm value and have incentives to search for private benefits at the
cost of minority shareholders. Furthermore, consistent with the evidence of previous empirical
literature, family managers are generally inferior to professional managers, providing that more
talented managers are more likely to be selected from the outside labor pool than within family
kinships (Morck, Stangeland, and Yeung (1998); Pérez-González (2001, 2006); Villalonga and
Amit (2006a)).
Intergenerational succession decision is one of the most observable corporate decisions in
family firms, and several factors can affect the transition decisions. Bennedsen, Nielsen, PerezGonzalez, and Wolfenzon (2007) shows that the gender of first born child is strongly correlated
with family CEO succession decisions. In addition, the financial cost could also influence the
succession decisions. Lagaras and Tsoutsoura (2015) finds that family firms has lower financial
cost and creditors value family ownership and control by explicitly imposing the presence of
family members to stay in the firm. Policy changes could also affect the family firm succession
decisions, as Tsoutsoura (2015) finds that inheritance taxes affect the probability of observing
family successions decisions and investment strategies. In this paper, we are attempting to see
how the pressure from the product market changes the top executive manager transitions.
As product market competition intensifies, incumbent firms generate less profit and become
more likely to be liquidated, thus reducing the managerial slack and increasing the cost of hiring
less qualified managers (Hart (1983); Schmidt (1997)). At the same time, the decreased mark-up
5
leaves less rewards for managers and lower the provision of effort. Raith (2001) shows mixed
results once allowing for the free entry of firms while changes the elasticity of substitution with
larger size of the market. We can think of a context where the board members in a family firm
hold annual meetings to decide the senior executive managerial position change. The decreased
output for survived firms driven by more substitutable products and larger markets changes the
mapping between the profit and the effort or the skill. The variation of product competition
changes the relative benefits and costs of both forcing current manager leave the firm and
appointing family members to key managerial positions. In this paper, we empirically test the
net effect.
Empirical work provides some evidence for the theoretical predictions. Several papers show
that the product market competition can substitute for internal governance. (Bertrand and Mullainathan (2003); Chhaochharia, Grinstein, Grullon, and Michaely (2012)). In addition,Giroud
and Mueller (2011) finds that companies in more competitive industries benefit less from corporate governance. Cuñat and Guadalupe (2005, 2009) show that product market competition
changes the pay structure of U.S. executives. The ambiguous theoretical analysis and contrasting
empirical research on family firms lead us to ask the question of how product market competition
affects the senior management turnover in family firms as a pre-step to answer deeper questions
such as do family firms respond efficiently to economic environment changes. Essentially, this
paper is a joint test of two hypothesis: 1) family transitions provide private benefits of control,
and 2) product market competition changes the mapping between the profits and the managerial
effort/skill.
3
Data and Sample Statistics
Our sample covers UK family firms in manufacturing sector from 2004 to 2014. In this section, we
briefly introduce the data collected from Orbis and Eurostats. Then we describe our definitions
of family firms and manager replacement. Sample summary is given at the end of this session.
The details of variable names and definition are available in Appendix 1.
6
3.1
Management and financial data
Orbis database of the Bureau van Dijk (BvD) provides comprehensive historical management
information on public and private firms all over the world. For each manager observation in the
database, we collect information such as age, gender, full name, level of responsibility as well
as appointment and resignation date of senior executive managers and board members. Firm
level financials from year 2004 to 2014 are obtained from the Fame database. The financial data
is matched with management data using the unique identifier generated by BvD. Based on the
individual senior executive manager information collected from Orbis, we next construct our
family firm sample and then define the senior manager departure and successions.
3.2
Family firm definition
The literature on family firm research uses a broad definition of family firms and this maybe the
potential source of diverging implications on performance of family business.1 We first define the
senior executive manager as a family member if by the time the manager is appointed; at least
one current or previous senior executive manager has the same family name. Considering the
nature of the research and the availability of management information, we classify a company
as a family firm if at least one family member is in the senior executive management team. A
founder manager is defined when the senior executive manager is appointed when the firm is
incorporated.2 A family founder is identified when at least two founders have the same family
name. There are some concerns about the “founder” definition. One is that two “founders” may
have the same family name accidentally and the other is that the managers are not usually the
owner of the firm. However, considering that the most firms in our sample are small, private
and young firms in the manufacturing sector, we can alleviate some of the concerns because the
smaller size reduces the probability that two senior managers have same family name accidentally
comparing to large and complicated conglomerates.
1
Other definitions include family ownership or control rights, or the inter-generation transfers.
The management record does not necessary dates back to the incorporation date. In this case, founder is
defined as managers who is the first to join the firm when earliest data is available.
2
7
3.3
Succession Definition
Succession may come in different forms, such as ownership, management and board members
replacement. In this paper, we focus on the senior executive management successions. For
each firm included in the database, we have detailed information of recorded senior executive
managers and board members. In our sample, more than 90% of the senior executive managers
are also board members. Since the information source varies from annual reports to informal
research and social networks, we do not have exact title of “Chief Executive Officer” or “Highest
Level Position” in most small and private companies. These companies usually only have less
than 5 reported senior managers, with one manager holding several positions at the same time.
For the majority of firms in our sample, there is usually no specification of one people at
the highest level and decisions are made collectively. We keep all senior executive managers
information if several managers are at the same level of position.
Before we define succession events, an accurate specification of resignation or departure is
required3 . A senior executive manager leaves the firm if the same person is not reappointed
to the same position within one year. Starting with the firm manager level data, we match
each family member departure date with the appointment dates and the family names of all
other senior executive managers in the same firm. We are able to find the best match for each
departing manager and define it as a “succession” whenever the following three criteria are met:
(1) the leaving manager must stayed in his position for at least one year; and (2) the leaving
manager cannot be matched to himself; and (3) the new manager is appointed around one year
of time when the previous manger leaves. When there are several candidate matches, we rank
the coming managers by the time gap, the position title then choose the one with the highest
rank. We keep all possible matches with the same rank. In case of no matches, we define the
departure with no replacement.
Again by checking the family names of the appointed manager, we separate the family
replacements from unrelated replacements. For the subsequent analysis, we measure the number
3
For simplicity we do not distinguish these two cases and call them departure events.
8
of total senior executive manager departures, the total departure of family senior executive
managers and the outcomes of family executive manager departures at firm-year level.
3.4
Competition Measurement
This paper focuses on the product market competition from foreign competitors. Following
Bertrand (2004) and Bloom, Sadun, and Van Reenen (2010), we use the import penetration as a
proxy for the product market competition from foreign markets. Import penetration is defined
as the ratio of import value over the summation of import and domestic production. The time
series of domestic production data at industry level for manufacturing sector is downloaded from
Eurostats Structural Business Statistics database (SBS). The international bilateral trade data
between UK and its 200 major trading countries from year 2004 to 2012 is collected from the
Comext database in Eurostat.
3.5
Sample statistics
Starting with 5,364,791 active UK firms in manufacturing sector from Orbis dataset, we exclude
firms incorporated before year 2004 and restrict to only family firms with available financial
data and in industries which we have industry-country level trade data. The time periods we
analyze is from year 2004 to 2014. Industry code changed from NACE rev1.1 to NACE Rev2
in 2007, so we follow Bena and Xu (2016) to link these two versions of code 4 . Overall, these
restrictions bring the number of our sample down to 13,478 companies. 10,788 family firms have
the family founder as senior executive managers (founder family firms) and 2,690 do not have
the family founder managers (nonfounder family firms).
The characteristics of family companies are showed in Table 1. The first two columns report
the summary statistics based on information of 13,478 family firms. The first part of the table
shows that the average number of senior managers is around 3, with 2 out of them from the
family and 1.6 being the family founders. In the second panel, we observe that the average
number of senior manager departure is about 0.06, and the family manager departure has an
4
The first 4 digit of the Prodcom code is the same as the NACE code throughout our sample. All the firms
report NACE Rev2 code in our sample.
9
average number of 0.04. 0.02 manager departures are not replaced with incoming new managers,
and only 0.006 family replacement after family manager departures happen each firm year on
average. In addition, there is a slightly higher number of unrelated replacements with the mean
of 0.008 and also higher standard deviation of 0.1. The average firm size is around £0.7 million
and the firm average age is around 18 years. We find that the age distribution is right skewed
because we restrict the companies to be incorporated before year 2004. The next two columns
show the summary statistics for the founder family firm sample, where the founder is still serving
as managers in the firm. Compared with the nonfounder family firm sample, founder family
firm sample has more family senior managers and higher number of unrelated replacements.
The founder family firms are larger and more matured firms.
In addition, the first Panel in Figure 1 shows the industry distribution of firms. We see
that no single industry has a portion of more than 0.3%, which indicates that all results are not
driven by a few industries in our sample. The second panel plots the time varying series of the
industry average import penetration level. We see that there is an increasing trend, with a high
growth rate before the crisis and a slightly drop after year 2008. The last panel shows the time
varying exchange rates. There is an appreciation of GBP around 2006 and 2012.
The summary statistics of import penetration and exchange rate index are included in
Table2. 10 industries are randomly selected from 188 industries from our sample. We can see
that there is a variation of import penetration and the import weighted exchange rates across
industries. At the bottom of the table, we show that the average of import penetration in our
panel sample is 0.34 and the exchange rate index average is 0.90. The standard deviations of
0.18 and 0.15 are large compared to the mean.
4
4.1
Empirical strategy
Hypothesis
Theoretical papers imply ambiguous effect of product market competition on providing managers’ incentives in a principal agent setting (Raith (2001); Schmidt (1997) etc).
10
When the product market competition increases, it changes the mapping between the profit
and the managerial effort. The reward of managerial effort can go up or down. In addition, the
firms are more likely to be liquidated, deciplining the managerial behaviour (Schmidt (1997)).
Raith (2001) argues that when we allow for the free entry into the market, there will be more
substituable products and the impact on profits and incentive payoff is mixed. Hart (1983) finds
when information is available, the assessment of manager’s effort are easier and also dicipines the
manager’s behaviour. The firm can improve the performance by either induce more managerial
effort or replace the managers with higher skilled ones, the logical relationship between product
market competition and managerial effort also applies to that with managerial skill.
The joint hypothesis we test are 1) product market competition changes the mapping between profit and managerial effort/skill; and 2) family managers do obtain the private benefit of
control. Under the hypothesis, when the product market competition is high, the private benefit of family control decreases. So we expect to see an increase in number of senior executive
manager departures.
It is not straightforward when we consider to differentiate the reward to skill channel from
the reward to effort channel through which the competition works.
If the leaving family manager is redundant (too poorly skilled), the competition reduces
the private benefit of family control. Therefore we expect to observe more family manager
departures without replacements. The decreased private benefit of control also makes family
successions less favourable. Providing the uncertain rewards of skill or effort with increasing
competition, it is reasonable to see either more or less frequent unrelated replacement.
For unrelated manager positions, if the skill of current manager is too low, then it is more
likely to be replaced with unrelated manager. Everything else equal, if the manager was not providing high effort, then we will expect the higher level competition affect all unrelated managers
same way. So there is no good reason to see a significant increase in unrelated replacement.
In general, we expect to observe that the board hire more outside members when the
monitoring costs are low (Boone, Field, Karpoff, and Raheja (2007)). The empirical tests is
11
aimed at evaluating the total effect of the joint hypothesis and shed some light on distinguishing
the skill from effort channel.
4.2
Empirical specifications
A simple and straightforward way to estimate the impact of product market competition on
family firm succession is to use the Ordinary Least Square regression. The baseline panel OLS
we run is in the form:
yf,t+1− t+2 = β0 + β1 Import Pen.i,t + xf,t β2 + Y earF Et + IndustryF Ei + f t
(1)
where f denotes a firm, i denotes an industry and t denotes time. yf,t+1− t+2 is the outcome
variable for firm f from time t + 1 to t + 2. In both the baseline OLS regressions and the
subsequent instrument variable regressions, variable yf,t+1− t+2 measures the dependent variables
in two years window respectively.
To investigate the effect of foreign competition on senior executive managerial position
changes, we run following two sets of regressions. In the first set of regression, we examine
whether more senior managers leave the firm when competition increases. Considering that
it takes time to make any change to current senior managerial positions, we look at the two
year window after the foreign competition changes. We then examine the number of family
manager departures and the unrelated manager departures separately. In the second set of
regressions, we focus on the outcomes of vacant managerial positions. We measure the total
number of managerial positions replaced by new family managers, F amilyReplacement and
other alternatives, which includes both the number of no replacement N oreplacement and the
number of unrelated replacement U nrelatedReplacement after the family manager departure.
All regressions are estimated over the full sample.
We denote by Import Pen.i,t the import penetration for industry i at time t. The coefficient
β1 is the main coefficient of interest. It measures the changes of senior executive manager
12
departures in response to 1 unit change of import penetration in the baseline regression. xf,t
denotes the control variables including firm age and size.
4.3
5
Endogeneity problem and Instrument variables
Equation 1 estimates the impact of foreign competition measured by import penetration on
senior executive managerial position changes. However, the common concern with the OLS
regressions is that the omitted variables may be correlated with the independent variable of
interest Import Pen.i,t . The endogeneity can be attributed to several reasons. The import penetration may be reversely determined by the executive manager position changes through the
shift of managers operating strategies. Furthermore, the import penetration for each industry
is the ratio of imports over the summation of imports and domestic production. For some of
our industries, due to privacy policy or the data collection restrictions, domestic production at
NACE 4 digit level is missing in some years, thus creating measurement error. In addition, the
concerns of endogeneity arise when the family firms anticipate the import penetration fluctuations, especially in the post-recession period when governments carried out stimulus packages
both in UK and within Europe. All these factors tend to bias our estimates down towards zero.
To address these concerns of endogeneity, we follow Bertrand (2004), and Cuñat and
Guadalupe (2009a) to implement import weighted exchange rates as instruments. For each
industry every year, we construct the industry specific exchange rates index as the bilateral
exchange rates between UK and other countries, with the weights being the import share of
total imports in the year 20026 . Specifically, we include both the current and lagged exchange
rate indices as our instruments considering that it usually takes time for import penetration to
have an impact. Thus the first stage regression at firm-year level is
5
We would like to add more firm level characteristics such as profit loss, turnover, leverage, etc. But most of
the firms are small firms and exempted from reporting financials to the Companies House.
6
The firm characteristic data is available from year 2004, and with the outcome estimation variable window of
2 years, we use the import share in year 2002
13
Import Pen.i,t = α0 +α1 Exch.Rateit +α2 Exch.Ratei,t−1 +xf,t α3 +Y earF Et +IndustryF Ei +ξf t
(2)
where Import Pen.i,t is the import penetration for industry i in year t. Exch. Ratei,t and
Exch. Ratei,t−1 are industry specific exchange rate indices in current and lagged year. xf,t is
defined as in equation 1. Year and industry fixed effects are controlled for and error terms are
clustered at the industry level.
The first stage regression results over different samples are listed in Table 3 in detail.
Column (1) to (3) report the first stage regression results over full sample, subsample with
family founders as senior executive manager in family firm and sub sample with non-founder
family senior managers in firm respectively. Starting with the full sample, we observe that 1%
of appreciation of pounds against the trading partner currencies is associated with a 0.79%
decrease of import penetration contemporary and increases the import penetration one year
later by 0.60%. The delayed positive reaction of import penetration to the appreciation of local
currency is a reflection of “J curve” effect in the related trade literature. Year fixed effect and
the industry fixed effect control for unobservable industry specific and time varying shocks. The
variation comes from the industry-time level exchange rates fluctuation. First stage F tests
result is reported and the they are all significant at 1% level. Hansen J test shows that the
over identification problem is not severer7 . The coefficients in column (2) and (3) have similar
patterns. In unreported tables, we also included the lagged two period exchange rates with
higher F test statistics. We only include the current and lagged exchange rates for keeping one
more years of observation.
The exclusion restriction assumption for the instrument is that the import weighted exchange rates only pick up the foreign market competition through the changes in industry level
import penetration. Import weighted exchange rates are less likely to be predicted by the companies. Using weights in year 2002 and the nominal exchange rates, we reduce the impact of
7
Notice that the p-value reported for subsample with transitions is larger compared to the previous two
columns, but this is not the main concern comparing to the exogeneity.
14
time varying import demand from different countries. The firms in our sample may have both
import and export trades, and the change in exchange rates may causes the change in export
openness (the export value over domestic production) at industry level. In fact, we run regressions of instrumental variables on export openness. Results are displayed in Table 2 from
column (4) to (6) over full sample, the subsample of founder firms and the subsample of nonfounder firms respectively. We see that the current and lagged industry level exchange rates are
poorly correlated with the export openness8 . These regression results reduce the possibility that
the exchange rates affect the managerial position changes through the indirect export openness
channel and adds to the validity of the instruments.
The second stage regression is:
yi,t+1
t+2
dPen. + xf,t β2 + +Y earF Et + IndustryF Ei + ξf t
= β0 + β1 Import
i,t
(3)
dPen. is the fitted value from equation 2, xf,t is the same set of control
where Import
i,t
variables as before. The coefficient β1 is of our main interest, which captures the change of
characteristics of senior managerial positions as the foreign competition increases through the
appreciation of Pounds.
5
5.1
Results
Senior executive manager departure
In this section we provide the baseline OLS and IV-2SLS regression results of senior executive
manager departure.
Table 4 presents the estimation results for the impact of foreign competition on the number of management departures. The regressions are estimated over the full sample with the
dependent variable Number of departures, the number of family departures and the number of
unrelated departures. OLS regression results suggest that the number of senior executive man8
The coefficients of regressions on export openness is larger in magnitude because export openness by definition
is not bounded to 1.
15
ager resignation does not appear to be significantly associated with the increase of the import
penetration. In contrast, when we instrument the import penetration with current and lagged
industry specific exchange rate indices, we observe that 1 unit increase in import penetration
results in a 0.29 more senior executive manager departures. The coefficient is significant at 1%
level with a standard error of 0.1022. The economic effect is also sizeable, 1 standard deviation
increase in import penetration causes the increase in number of departures by 0.052 (0.287*0.182)
in the future two years, which is approximately 44% (0.052/(2*0.059))of the sample average.
When we examine the number of family departures and unrelated departures more closely,
we find that OLS regressions imply a positive correlation of family departures with foreign
competition whereas the number of unrelated departures is negatively correlated with the competition. IV-2SLS regression in column (4) shows that 1 unit increase of import penetration
causes 0.21 increase in number of family manager departures, significant at 5% level. The
coefficient is also economically significant, 1 standard deviation increase in import penetration leads to 0.038(0.0182*0.207) more family manager leave the firm in the future two years,
49%(0.038/(2*0.038)) of the sample average. However, the 0.08 increase in the number of unrelated manager departures is not significant.
As we can see from Table 4 that the OLS estimates are substantially biased towards zero
because of the correlation of import penetration measurement of competition with the omitted
variables. In addition, larger firms and older firms appear to significantly increase the number
of decisions to fire the manager given the import penetration level, though at a much smaller
magnitude (0.015 and 0.001 in column 2, Table4).
Overall, Table 4 add evidence to support the hypothesis that the competition from foreign
markets reduces the private benefit of family control and lead to more senior managers, in
particular the family managers leave the position (no matter if the outgoing manager is replaced
by a new manager or not). The more significant increase in family managers departure suggesting
that they are less qualified and redundant.
16
5.2
Outcomes of family manager departures
In Table 5, we further investigate the decision after the family manager leave the position.
There are three outcomes after the family manager leave the position: not replaced by any
new managers, replaced by family managers or replaced by unrelated managers. We report the
regressions on these outcomes separately in column (1) and (2), column (3) and (4), column
(5) and (6). All the regression results are estimated over full sample at firm year level. The
dependent variables are calculated in forward two year window as before.
The first two columns show us that in face of high foreign competition, firms choose not to
replace the departing senior executive managers in the coming two years. 1 standard deviation
increase of import penetration leads to 0.032 (0.0317*0.182) more family managers leave the
firm without replacement in the future two years, accounts for nearly 68% (0.032/(2*0.023)) of
the sample average. Although there is a positive increase of family successions and unrelated
successions in column (4) and column (6) of 0.0298 and 0.0191 respectively, the coefficients
are not statistically significant and the magnitude is small compared to the no replacement
coefficient.
The results in Table 4 and 5 suggest that the increased import penetration causes an increase
in the number of senior executive manager resignations. Furthermore, we do not observe a
significant increase in either family succession nor unrelated successions. Family firms choose
not to replace the leaving family manager positions more often as response to the increasing
competition. The result is statistically significant and at 5% and has economically significant
interpretations.
5.3
Outcomes of unrelated manager departures
We analyse the decision after the unrelated manager leave the position in Table 6. Similar to the
case when family manager leave the firm, there are three outcomes after the unrelated manager
leave the firm: not replaced by any new managers, replaced by unrelated managers or replaced
by family managers. We report the regressions on these outcomes separately in columns (1) and
17
(2), (3) and (4), (5) and (6). All the regression results are estimated over full sample at firm
year level. The dependent variables are calculated in forward two year window.
The first two columns show us that intense competition results in an increasing number of
no replacement after unrelated senior executive managers departures in the coming two years,
but the coefficient is not statistically significant. The next two columns report that there is a
significant increase in number of unrelated managers successions at 5% level. Economically, 1
standard deviation of increase in import penetration results in a 0.0083 (0.0457*0.0182) increase
in the unrelated succession of unrelated managers in the future two years. This accounts for
about 105% (0.0083/(0.004*2))of the sample average. We also observe an increase of family
successions in column (6), however, the coefficient is not statistically significant.
Overall, the results in Table 6 show that the foreign competition causes an increase in the
number of unrelated managers leave the firm and replaced by the unrelated managers in the
future two years. Together with the results we displayed in Table 5, we conjecture that with more
competition from foreign market, redundant family managers are fired and important unrelated
manager positions are replaced with more capable outside managers. These results implies that
the skill channel fits the observed regression results better.
5.4
Do family founder firms behave differently?
Founder CEO run firms are more efficient than other family firms based on findings from previous
literature. Villalonga and Amit (2006b) finds that founder CEOs create value inside the family
firms though in general the family firms are underperformed compared with other firms. The idea
is that the principal agent conflicts are more costly in nonfamily firms, whereas the nonfounder
family firms faces more conflicts between family member and outsiders.
In this section, we test whether family firms with the family founder managers react differently to increase foreign competition than nonfounder family firms in our sample. We examine
the regression results over sub samples with and without family founder as senior executive
manager respectively. The regression we run are similar to the previous ones over the full
sample.
18
Table 7 show IV estimation results using founder family firms and nonfounder family firms.
The dependent variable in first two columns is the total number of senior manager departures
in coming two years window. The next two columns show the impact of competition on the
number of family manager departures and the regression on the number of unrelated departures
are reported in the last two columns. A direct comparison between Table4 and Table7 shows
that the results for the full sample holds for family founder firms. The significant increases in
both total number of manager departure and family manager departure are reported in column
(1) and (3), with 0.049 (0.27*0.1815) and 0.039 (0.21*0.1815) increase respectively in face of 1
standard deviation increase in the import penetration. Contrarily, for the nonfounder family
firms, we do not observe a significant increase in family departure and they tend to keep the
current managerial team structure stable.
Similar to 7, Table 8 shows the regression results with dependent variables number of no
replacements, number of family successions, and the number of unrelated successions. Not
surprisingly, we observe that 1 standard deviation increase in import penetration causes 0.034
(0.1883*0.182) more family managers departures without replacement. The result is significant
at 5% level. Table 9 column (3) also reports a significant increase in the number of unrelated
replacement of unrelated manager departures caused by intense competition. However, we
observe insignificant estimations in almost all regressions within nonfounder family firms, with
only number of manager departures increasing significantly at 10%.
To conclude, the main results we showed in the previous section remains in the subsample of
founder family firms, however, the nonfounder family firms seem to be more silent and reluctant
to change the current manager team structure. The findings support the arguments in Villalonga
and Amit (2006b) that founder adds valuable management skill to the firm while descendent
CEO destroys the firm value.
5.5
Merge and Acquisitions after the managerial position changes
To conclude so far, we found that more competition causes more family managers departing
the firm without replacement, unrelated managers are replaced with unrelated managers. A
19
natural question to ask is weather the change of management team has any economic impact on
the firm. One interesting aspect to look at is the probability of being acquired by other firms.
Under the hypothesis, combined with the previous findings that managerial position changes in
response to the intense competition, family firms which fire more family managers and replace
them with unrelated managers may face more pressure to be acquired.
We then complement the family firm sample with the merge and acquisition events documented in the Zephyr database in Orbis. The variable Acquired is the dummy variable takes
the value of 1 if the firm is acquired by other firms in the following two years. In total, there
are 128 family firms being acquired in our sample from year 2004 to 2014. The OLS regression
results on acquisition are displayed in Table10.
We find that the number of manager departures, especially the family manager departure
is positively correlated with the probability of being acquired in future two years. Further more,
column (4) and (6) show that family firms acquisition is positively associated with Unrelated
successions after the family and unrelated manager departures. The interpretation could be
that the competition forces the underperformed firms to fire unqualified managers and those are
not able to survive then are acquired by other firms. Or it could be that family firms that are
financial constrained choose to sell the firm out and family managers are exiting and replaced
by unrelated managers.
6
Robustness check
We include both the current and lagged exchange rate index as our instrumental variables of
import penetration. However, because of the “J curve” effect, the opposite coefficients of the two
instruments presented in the first stage regressions may be concerning. Therefore, it is worth
checking the reduced form regressions.
We report all the reduced form regressions for the full sample, the founder family firm
sample and the nonfounder family firm sample from Table11 to Table13. We observe some
mixed results produced in Table11, the lagged exchange rates have positive and significant effect
20
on the number of manager departures, the family manager departures. The appreciation of
lagged exchange rates causes more family managers to be replaced by the unrelated managers or
without replacement after they leave the firm. Unrelated managers are more likely to be replaced
by the unrelated managers. The current exchange rate works in the opposite way though, the
coefficients are mostly negatively significant.
When we only include the subsample which includes founder running family firms, we
observe similar patterns in estimated coefficients in Table12. Comparing the magnitude and
the coefficients in front of lagged and current exchange rates, we find that the net effect on
total number of manager departures, the number of family manager departures are positive
and significant. Similar to the previous findings, the net effect on unrelated replacement after
unrelated leave the firm is positive significant. When the family members leave the firm, the
reduced form regression suggests that the lagged appreciation of Pounds lead to an increase of
unrelated succession.9
7
Conclusion
In this paper, we use the detailed senior manager information of UK family firms in manufacture
sector to examine the impact of foreign competition on family firm executive managerial position
changes. In fact, we examine the specific channel, the foreign competition through which the
market competition works. We measure the foreign competition using the import penetration
at the industry level and instrument it with current and lagged import weighted exchange rates.
We find that increasing product market competition from foreign countries causes a significant
increase in the number of senior manager resignations in family firms, especially the number of
family senior manager resignations. We then examine the outcomes after the manager departures
and show that foreign competition results in an increase number of family manger departures
without replacement and the number of unrelated replacement after unrelated manager leave
9
In the unreported tables, we also run regressions with subsample from 2004 to 2013, the IV-2SLS results
have same pattern with the regression results displayed in the main tables. In reduced form regressions, the net
effects of current and lagged exchange rate index are significantly positive when dependent variables are number
of departures, family departures, no replacement after family departure and unrelated succession after unrelated
departure.
21
the firm. Moreover, we find that the family firm is more likely to be acquired in two years after
more managers are replaced with unrelated managers. Overall, these results suggest that we
can not reject the joint hypothesis that 1) within family transitions provide private benefits of
family control and 2) the product market competition changes the reward of managerial effort
and skill. While the paper does not distinguish the effort channel from the skill channel, the
findings are more in favour of the skill channel. In the subsample tests, we show that family
firms without founders are more silent in face of foreign competition. Our results highlight that
foreign competition has a statistically significant and sizeable causal effect on senior manager
position changes in family firms. The paper adds an important economic force, the foreign
competition to the existing factors such as firm characteristics, institutional environments and
the tax policies affecting the family firm succession plans. Future works are needed to investigate
the mechanism behind the driving force of family firm succession decisions.
22
References
Anderson, Ronald C., and David M. Reeb, 2003, Founding-family ownership and firm performance: Evidence from the s&p 500, The Journal of Finance 58, pp. 1301–1328.
Bena, Jan, and Ting Xu, 2016, Competition and ownership structure of closely-held firms, Available at SSRN: http://ssrn.com/abstract=2356526 or http://dx.doi.org/10.2139/ssrn.2356526.
Bennedsen, Morten, Kasper Meisner Nielsen, Francisco Perez-Gonzalez, and Daniel Wolfenzon,
2007, Inside the family firm: The role of families in succession decisions and performance, The
Quarterly Journal of Economics 122, 647–691.
Bertrand, Marianne, 2004, From the invisible handshake to the invisible hand? how import
competition changes the employment relationship, Journal of Labor Economics 22, 723–765.
, and Sendhil Mullainathan, 2003, Enjoying the quiet life? corporate governance and
managerial preferences, Journal of Political Economy 111, 1043–1075.
Bloom, Nicholas, Mirko Draca, and John Van Reenen, 2016, Trade induced technical change?
the impact of chinese imports on innovation, it and productivity, The Review of Economic
Studies 83, 87–117.
Bloom, Nicholas, Carol Propper, Stephan Seiler, and John Van Reenen, 2015, The impact of
competition on management quality: evidence from public hospitals, The Review of Economic
Studies p. rdu045.
Bloom, Nicholas, Raffaella Sadun, and John Van Reenen, 2010, Does product market competition lead firms to decentralize?, Rock Center for Corporate Governance at Stanford University
Working Paper.
Bolton, Patrick, Von Thadden, et al., 1998, Blocks, liquidity, and corporate control, The Journal
of Finance 53, 1–25.
Boone, Audra L., Laura Casares Field, Jonathan M. Karpoff, and Charu G. Raheja, 2007,
The determinants of corporate board size and composition: An empirical analysis, Journal of
Financial Economics 85, 66 – 101.
23
Chhaochharia, Vidhi, Yaniv Grinstein, Gustavo Grullon, and Roni Michaely, 2012, Product
market competition and internal governance: Evidence from the sarbanes oxley act, Johnson
School Research Paper Series.
Claessens, Stijn, Simeon Djankov, and Larry H.P Lang, 2000, The separation of ownership and
control in east asian corporations, Journal of Financial Economics 58, 81 – 112 Special Issue
on International Corporate Governance.
Cuñat, Vicente, and Maria Guadalupe, 2005, How does product market competition shape
incentive contracts?, Journal of the European Economic Association pp. 1058–1082.
, 2009, Globalization and the provision of incentives inside the firm: The effect of foreign
competition, Journal of Labor Economics 27, 179–212.
Ellul, Andrew, Marco Pagano, and Fausto Panunzi, 2010, Inheritance law and investment in
family firms, The American Economic Review pp. 2414–2450.
Faccio, Mara, and Larry HP Lang, 2002, The ultimate ownership of western european corporations, Journal of financial economics 65, 365–395.
Giroud, Xavier, and Holger M Mueller, 2010, Does corporate governance matter in competitive
industries?, Journal of Financial Economics 95, 312–331.
, 2011, Corporate governance, product market competition, and equity prices, The Journal of Finance 66, 563–600.
Guadalupe, Maria, and Francisco Pérez-González, 2010, Competition and private benefits of
control, in AFA 2007 Chicago Meetings Paper.
Hart, Oliver D, 1983, The market mechanism as an incentive scheme, The Bell Journal of
Economics pp. 366–382.
Kadyrzhanova, Dalida, and Matthew Rhodes-Kropf, 2011, Concentrating on governance, The
Journal of Finance 66, 1649–1685.
Lagaras, Spyridon, and Margarita Tsoutsoura, 2015, Family control and the cost of debt: Evidence from the great recession, Fama-Miller Working Paper pp. 15–14.
24
Morck, Randall K., David A. Stangeland, and Bernard Yeung, 1998, Inherited wealth, corporate
control and economic growth: The canadian disease, Working Paper 6814 National Bureau of
Economic Research.
Pérez-González, Francisco, 2001, Does inherited control hurt firm performance, Unpublished
manuscript. Columbia University.
, 2006, Inherited control and firm performance, The American economic review pp.
1559–1588.
Porta, Rafael, Florencio Lopez-de Silanes, and Andrei Shleifer, 1999, Corporate ownership
around the world, The journal of finance 54, 471–517.
Raith, Michael, 2001, Competition, risk and managerial incentives, Available at SSRN 262648.
Schmidt, Klaus M, 1997, Managerial incentives and product market competition, The Review
of Economic Studies 64, 191–213.
Shleifer, Andrei, and Robert W Vishny, 1986, Large shareholders and corporate control, The
Journal of Political Economy pp. 461–488.
Tsoutsoura, Margarita, 2015, The effect of succession taxes on family firm investment: Evidence
from a natural experiment, The Journal of Finance 70, 649–688.
Villalonga, Belen, and Raphael Amit, 2006a, How do family ownership, control and management
affect firm value?, Journal of financial Economics 80, 385–417.
, 2006b, How do family ownership, control and management affect firm value?, Journal
of Financial Economics 80, 385 – 417.
25
Figure 1
0
.001
Density
.002
.003
Industry Distribution Full Sample
500
1000
1500
2000
2500
NACE Rev. 2Core code (4 digits)
3000
3500
.36
Average Import Penetration
.38
.4
.42
Time varying average Import Penetration across industry
2004
2005
2006
2007
2008
time
2009
2010
2011
2012
.8
Average exchange rates
.85
.9
.95
1
Time varying average of exchange rates across industries
2004
2006
2008
time
2010
2012
18.029
0.718
0.023
0.006
0.008
0.030
0.004
0.002
16.715
1.313
0.166
0.080
0.107
0.184
0.063
0.044
0.282
0.226
0.153
1.284
0.949
0.991
17.725
0.709
0.023
0.006
0.009
0.022
0.003
0.001
0.057
0.041
0.016
2.908
2.365
2.008
17.052
1.319
0.166
0.084
0.111
0.161
0.058
0.039
0.278
0.235
0.135
1.262
0.908
0.647
Founder Sample
Mean
S.d
19.236
0.755
0.023
0.004
0.007
0.060
0.006
0.003
0.067
0.028
0.039
2.962
1.542
0
15.251
1.287
0.167
0.060
0.092
0.256
0.081
0.059
0.296
0.184
0.208
1.367
0.814
0
Nonfounder Sample
Mean
S.d
Num. of firms
13,478
10,788
2,690
Notes: All statistics are reported for family firms with at least one year financial data available from year 2004 to 2014. Further
restrictions requires the firms are operating in industries with trade data non-missing and total number of senior managers in base
year is less than 10. The first two columns report the statistics over full sample of firms, and the next two columns describe the sub
sample of firms with family founder and the last two columns corresponds to subsample with non founders. The reported numbers
of observation varies across variables due to the availability. Assets refers to total assets in the balance sheet by the end of year
2004, in million £.
Age in base year
Size (Assets in mil)
Num. of unrelated manager departure
Num. of Family manager departure
No replacement
family succession
unrelated succession
No replacement
unrelated succession
family succession
0.059
0.038
0.021
Num. of manager departures
Num. of Family manager departures
Num. of Unrelated manager departures
with
with
with
with
with
with
2.919
2.199
1.604
Num. of senior managers
Num. of family senior managers
Num. of family founder senior managers
Full sample
Mean
S.d
Table 1: Summary Statistics: Dependent Varibles
26
27
Table 2: Summary Statistics: Independent variable
selected industries
2042
2052
2053
2059
2060
2110
2120
2211
2219
2221
Aggregated
Import penetration
mean
s.d.
0.453007
0.073992
0.199135
0.024927
0.493558
0.015957
0.436567
0.037978
0.501579
0.108124
0.57625
0.189322
0.461797
0.12776
0.467761
0.049425
0.381432
0.047258
0.315675
0.027678
0.341456
0.181531
Exchange
mean
0.875388
0.876904
0.869143
0.909582
0.877722
0.867873
0.875264
0.851707
0.889449
0.847686
0.902708
rate
s.d.
0.096321
0.099103
0.093522
0.105368
0.089384
0.118389
0.096247
0.105352
0.11349
0.097001
0.153075
Note: All statistics are reported for family firms with at least one year financial data available
from year 2004 to 2014. Further restrictions requires the firms are operating in industries with
trade data non-missing and total number of senior managers in base year is less than 10. In
this table, 10 industries are randomly selected out of 118 industries in our sample. For each
industry, the mean and the standard deviation of the import penetration and its exchange rate
indices are displayed. Import penetration is defined as the ratio of import value over the sum of
import and domestic production. Exchange rates is defined as weighted exchange rates between
trading country and UK, with import shares in 2002 as the weights. Import penetration and
the exchange rates are measured at the industry level.
28
Table 3: First stage regressions
(1)
Lagged Exch. Rates
Exch. Rates
Total Assets
Age
Import Penetration
(2)
(3)
0.6110**
(0.2485)
-0.7895***
(0.1713)
0.0000
(0.0001)
-0.0000
(0.0000)
0.6140**
(0.2466)
-0.7834***
(0.1664)
0.0001
(0.0001)
-0.0000
(0.0000)
0.5955**
(0.2600)
-0.8135***
(0.1961)
-0.0001
(0.0001)
0.0000
(0.0000)
(1)
1.5291
(1.5267)
-2.1379*
(1.1814)
0.0009*
(0.0005)
-0.0000**
(0.0000)
Export Openess
(2)
1.3852
(1.4634)
-2.0326*
(1.1271)
0.0009*
(0.0005)
-0.0000**
(0.0000)
(3)
2.1409
(1.8861)
-2.5791*
(1.4485)
0.0009
(0.0010)
-0.0000
(0.0000)
Num. of Obs
117,051
93,598
23,453
116,138
92,901
23,237
Num. of Industry
177
177
153
171
171
150
Year FE
YES
YES
YES
YES
YES
YES
Industry FE
YES
YES
YES
YES
YES
YES
R squared
0.0699
0.0692
0.0724
0.0102
0.0099
0.0119
F test
11.0583*** 11.5811***
8.8138***
2.0563
2.1087
1.8453
Sample
Full
Founder
Nonfounder
Full
Founder Nonfounder
2nd Stage Dependent variable: outcome in future 2 year window
Notes: The table shows the first stage regression results of equation 2 with the dependent
variable of import penetration. Column (1), (2) and (3) presents the results with different
sample. Column (1) includes the full sample, column (2) covers the subsample with family
founders as senior managers in the base year, and column (3) is estimated using subsample
with family firm without family founders in senior manager team. Import Penetration is import
divided by import plus domestic production at four digit industry level. Column (4) to (6) show
the exclusion restriction tests over different samples, where the dependent variable is the export
openness. Export openness is defined as export value divided by the domestic production value.
Industry specific exchange rates is the bilateral exchange rates between UK and its main trade
partners, weighted by the import share from each country in the base year. Assets is the total
assets in the balance sheet each year in millions. All regressions control for year and industry
fixed effects and are estimated at firm-year level. * significant at 10%; ** at 5%; *** at 1%.
Standard errors (in brackets) are clustered at industry level for all the regressions.
-0.0045
(0.0268)
0.0150***
(0.0025)
0.0010***
(0.0001)
0.2870***
(0.1022)
0.0150***
(0.0025)
0.0010***
(0.0001)
0.0030
(0.0232)
0.0023**
(0.0010)
0.0007***
(0.0001)
0.2070**
(0.1040)
0.0023**
(0.0009)
0.0007***
(0.0001)
Number of family departures
OLS
IV-2SLS
(3)
(4)
-0.0075
(0.0157)
0.0128***
(0.0020)
0.0003***
(0.0001)
0.0800
(0.0530)
0.0128***
(0.0020)
0.0003***
(0.0001)
Number of unrelated departures
OLS
IV-2SLS
(5)
(6)
Number of Obs.
117,051
117,051
117,051
117,051
117,051
117,051
R-squared
0.005
0.004
0.002
0.001
0.007
0.007
Num. of Industry
177
177
177
177
177
177
Year FE
YES
YES
YES
YES
YES
YES
Industry FE
YES
YES
YES
YES
YES
YES
Sample
Full
Full
Full
Full
Full
Full
Notes: The table shows the ols and 2SLS regression results. The dependent variable is the number of total senior manager departures in two years window after the competition shock in first two columns. Column (3) and (4) present the regression on number
of family manager departures. The last two columns show the regression results of unrelated senior manager departures in two
years forward window. Column (1), (3) and (5) present the OLS regression results; column(2),(4) and (6) present the baseline
IV-2SLS estimations with import penetration instrumented by current and lagged industry specific exchange rates respectively.
Industry exchange rate is weighted average of exchange rates at the industry level, where the weight is the import share from
each country in 2002. The independent variable import penetration is the value of import divided by the sum of import and the
domestic production. Assets is the total assets in the balance sheet each year in million £. All regressions control for year and
industry fixed effects and are estimated at firm-year level over the full panel. * significant at 10%; ** at 5%; *** at 1%. Standard
errors (in brackets) are clustered at industry level for all the regressions.
Age
Total Assets
Import Penetration
Number of departures
OLS
IV-2SLS
(1)
(2)
Table 4: Competition and Manager Departures
29
-0.0119
(0.0192)
0.0006
(0.0008)
0.0004***
(0.0001)
0.1741**
(0.0876)
0.0006
(0.0008)
0.0004***
(0.0001)
0.0067
(0.0062)
0.0002
(0.0004)
0.0002***
(0.0000)
0.0298
(0.0272)
0.0002
(0.0004)
0.0002***
(0.0000)
Family Replacement
OLS
IV-2SLS
(3)
(4)
0.0098
(0.0082)
0.0015***
(0.0005)
0.0001***
(0.0000)
0.0191
(0.0319)
0.0015***
(0.0005)
0.0001***
(0.0000)
Unrelated Replacement
OLS
IV-2SLS
(5)
(6)
Number of Obs.
117,051
117,051
117,051
117,051
117,051
117,051
Adjusted R-squared
0.001
-0.000
0.001
0.001
0.000
0.000
Number of Industry
177
177
177
177
177
177
Year FE
YES
YES
YES
YES
YES
YES
Industry FE
YES
YES
YES
YES
YES
YES
Sample
Full
Full
Full
Full
Full
Full
Notes: The table shows the OLS and IV-2SLS regression results on different outcomes of family manager departure. The dependent
variable is the number of family departures with no replacement in first two column; the number of family successions in column
(3) and (4), the number of unrelated successions in last two columns. All the dependent variable numbers are calculated in two
years forward window. Column (1),(3) and (5) present the OLS regression results; column(2),(4) and (6) present the baseline
IV-2SLS estimations with import penetration instrumented by current and lagged industry specific exchange indices respectively.
Industry exchange rate is weighted average of exchange rates at the industry level, where the weight is the import share from
each country in 2002. The independent variable import penetration is the value of import divided by the sum of import and the
domestic production. Assets is the total assets in the balance sheet each year in million £. All regressions control for year and
industry fixed effects and are estimated at firm-year level. * significant at 10%; ** at 5%; *** at 1%. Standard errors (in brackets)
are clustered at industry level for all the regressions.
Age
Total Assets
Import Penetration
No replacement
OLS
IV-2SLS
(1)
(2)
Table 5: Competition and Outcomes of Family Departures
30
-0.0083
(0.0141)
0.0086***
(0.0013)
0.0002***
(0.0000)
0.0254
(0.0370)
0.0086***
(0.0013)
0.0002***
(0.0000)
0.0013
(0.0081)
0.0031***
(0.0006)
0.0000
(0.0000)
0.0457**
(0.0210)
0.0031***
(0.0006)
0.0000
(0.0000)
Unrelated Replacement
OLS
IV-2SLS
(3)
(4)
-0.0005
(0.0052)
0.0011***
(0.0003)
0.0000***
(0.0000)
0.0089
(0.0133)
0.0011***
(0.0003)
0.0000***
(0.0000)
Family Replacement
OLS
IV-2SLS
(5)
(6)
Number of Obs.
117,051
117,051
117,051
117,051
117,051
117,051
Adjusted R-squared
0.005
0.005
0.002
0.002
0.001
0.001
Number of industry
177
177
177
177
177
177
Year FE
YES
YES
YES
YES
YES
YES
Industry FE
YES
YES
YES
YES
YES
YES
Sample
Full
Full
Full
Full
Full
Full
Notes: The table shows the OLS and IV-2SLS regression results on different outcomes of unrelated manager departure. The
dependent variable is the number of unrelated departures with no replacement in first two column; the number of unrelated
successions in column (3) and (4), the number of family successions in last two columns. All the dependent variable numbers are
calculated in two years forward window. Column (1),(3) and (5) present the OLS regression results; column(2),(4) and (6) present
the baseline IV-2SLS estimations with import penetration instrumented by current and lagged industry specific exchange indices
respectively. Industry exchange rate is weighted average of exchange rates at the industry level, where the weight is the import
share from each country in 2002. The independent variable import penetration is the value of import divided by the sum of import
and the domestic production. Assets is the total assets in the balance sheet each year in million £. All regressions control for year
and industry fixed effects and are estimated at firm-year level. * significant at 10%; ** at 5%; *** at 1%. Standard errors (in
brackets) are clustered at industry level for all the regressions.
Age
Total Assets
Import Penetration
No replacement
OLS
IV-2SLS
(1)
(2)
Table 6: Competition and Outcomes of Unrelated Departures
31
0.2698**
(0.1060)
0.0127***
(0.0020)
0.0013***
(0.0001)
0.4143*
(0.2240)
0.0228***
(0.0080)
-0.0004
(0.0003)
0.2130**
(0.0968)
0.0019*
(0.0010)
0.0010***
(0.0001)
0.2082
(0.2031)
0.0032
(0.0021)
-0.0005***
(0.0001)
Number of family departures
(3)
(4)
0.0568
(0.0401)
0.0108***
(0.0015)
0.0003***
(0.0001)
0.2061
(0.1726)
0.0196***
(0.0063)
0.0001
(0.0002)
Number of non-family departures
(5)
(6)
Number of Obs.
93,598
23,453
93,598
23,453
93,598
23,453
R-squared
0.005
0.004
0.002
-0.000
0.007
0.007
Number of Industry
177
153
177
153
177
153
Year FE
YES
YES
YES
YES
YES
YES
Industry FE
YES
YES
YES
YES
YES
YES
Sample
Founder
Nonfounder
Founder
Nonfounder
Founder
Nonfounder
Notes: The table shows the 2SLS regression results over sample with and without family founder as senior manager in base year
2004. The dependent variable is the number of total senior manager departures in two years window after the competition shock
in first two columns. The last two columns show the regression results of family senior manager departures in two years forward
window. All the columns report IV-2SLS estimations with import penetration instrumented by current and lagged industry specific
exchange rates respectively. Industry exchange rate is weighted average of exchange rates at the industry level, where the weight
is the import share from each country in 2002. The independent variable import penetration is the value of import divided by the
sum of import and the domestic production. Assets is the total assets in the balance sheet each year in million £. All regressions
control for year and industry fixed effects and are estimated at firm-year level. * significant at 10%; ** at 5%; *** at 1%. Standard
errors (in brackets) are clustered at industry level for all the regressions.
Age
Total Assets
Import Penetration
Number of departures
(1)
(2)
Table 7: Competition and Manager Departures-Founder firm vs. Nonfounder firms
32
0.1883**
(0.0857)
-0.0000
(0.0006)
0.0005***
(0.0001)
0.1400
(0.1557)
0.0027
(0.0026)
-0.0001
(0.0001)
0.0111
(0.0381)
0.0017***
(0.0006)
0.0002***
(0.0000)
0.0545
(0.0481)
0.0007
(0.0009)
-0.0001
(0.0001)
Family Replacement
(3)
(4)
0.0300
(0.0296)
0.0002
(0.0005)
0.0003***
(0.0000)
0.0247
(0.0474)
-0.0001
(0.0006)
-0.0001*
(0.0001)
Family to nonrelated Replacement
(5)
(6)
Number of Obs.
93,598
23,453
93,598
23,453
93,598
23,453
R-squared
-0.000
-0.001
0.001
0.000
0.001
0.001
Number of Industry
177
153
177
153
177
153
Year FE
YES
YES
YES
YES
YES
YES
Industry FE
YES
YES
YES
YES
YES
YES
Sample
Founder
Nonfounder
Founder
Nonfounder
Founder
Nonfounder
Notes: The table shows the IV-2SLS regression results. The dependent variables are the number of family successions in first two
coloumn, the number of non family-successions in column (3) and (4), the number of no replacement in column (5) and (6) and
the number of unrelated successions in last two columns. All the dependent vaiable numbers are calculated in two years forward
window. Column (1),(3),(5) and (7) present the regression results over subsample of founder family firms. Column(2),(4),(6) and
(8) present the regression results over subsample witout family founder firms. The import penetration instrumented by current
and lagged industry specific exchange indices respectively. Industry exchange rate is weighted average of exchange rates at the
industry level, where the weight is the import share from each country in 2002. The independent variable import penetration is
the value of import divided by the sum of import and the domestic production. Assets is the total assets in the balance sheet each
year in million £. All regressions control for year and industry fixed effects and are estimated at firm-year level. * significant at
10%; ** at 5%; *** at 1%. Standard errors (in brackets) are clustered at industry level for all the regressions.
Age
Total Assets
Import Penetration
No replacement
(1)
(2)
Table 8: Competition and Outcomes of Family Departures-Founder firm vs. Nonfounder firms
33
0.0019
(0.0313)
0.0073***
(0.0011)
0.0002***
(0.0000)
0.1473
(0.1421)
0.0130***
(0.0047)
0.0002
(0.0002)
0.0344**
(0.0163)
0.0027***
(0.0006)
0.0001**
(0.0000)
0.0961
(0.0668)
0.0045***
(0.0017)
-0.0001**
(0.0001)
Unrelated Replacement
(3)
(4)
0.0205
(0.0151)
0.0008*
(0.0004)
0.0000**
(0.0000)
-0.0374
(0.0313)
0.0021*
(0.0011)
0.0001
(0.0000)
Nonfamily to family Replacement
(5)
(6)
Number of Obs.
93,598
23,453
93,598
23,453
93,598
23,453
R-squared
0.004
0.005
0.002
0.002
0.000
0.001
Number of Industry
177
153
177
153
177
153
Year FE
YES
YES
YES
YES
YES
YES
Industry FE
YES
YES
YES
YES
YES
YES
Sample
Founder
Nonfounder
Founder
Nonfounder
Founder
Nonfounder
Notes: The table shows the IV-2SLS regression results. The dependent variables are the number of family successions in first two
column, the number of non family-successions in column (3) and (4), the number of no replacement in column (5) and (6) and
the number of unrelated successions in last two columns. All the dependent vaiable numbers are calculated in two years forward
window. Column (1),(3),(5) and (7) present the regression results over subsample of founder family firms. Column(2),(4),(6) and
(8) present the regression results over subsample witout family founder firms. The import penetration instrumented by current
and lagged industry specific exchange indices respectively. Industry exchange rate is weighted average of exchange rates at the
industry level, where the weight is the import share from each country in 2002. The independent variable import penetration is
the value of import divided by the sum of import and the domestic production. Assets is the total assets in the balance sheet each
year in million £. All regressions control for year and industry fixed effects and are estimated at firm-year level. * significant at
10%; ** at 5%; *** at 1%. Standard errors (in brackets) are clustered at industry level for all the regressions.
Age
Total Assets
Import Penetration
No replacement
(1)
(2)
Table 9: Competition and Outcomes of Unrelated Departures-Founder firm vs. Nonfounder firms
34
0.0007***
(0.0002)
-0.0000**
(0.0000)
(1)
0.0014**
(0.0006)
0.0007***
(0.0002)
-0.0000**
(0.0000)
0.0012*
(0.0007)
(2)
0.0007***
(0.0002)
-0.0000*
(0.0000)
0.0001
(0.0014)
0.0007***
(0.0002)
-0.0000**
(0.0000)
0.0014*
(0.0009)
0.0007***
(0.0002)
-0.0000**
(0.0000)
0.0118
(0.0082)
0.0069**
(0.0034)
0.0007***
(0.0002)
-0.0000**
(0.0000)
(6)
0.0007***
(0.0002)
-0.0000*
(0.0000)
(7)
Observations
117,051
117,051
117,051
117,051
117,051
117,051
117,051
R-squared
0.001
0.001
0.001
0.001
0.001
0.001
0.001
Num. of Industries
177
177
177
177
177
177
177
Year FE
YES
YES
YES
YES
YES
YES
YES
Industry FE
YES
YES
YES
YES
YES
YES
YES
Sample
full
full
full
full
full
full
full
Notes: The table shows the OLS regression results. The dependent variable is Acquired, a dummy variable which takes the
number of 1 if the firm is acquired in the future two years window. The independent variables are number of manager departures,
number of family manager departures, the number of family or unrelated replacement after family or unrelated manager departures
respectively. Assets is the total assets in the balance sheet each year in million £. All regressions control for year and industry
fixed effects and are estimated at firm-year level. * significant at 10%; ** at 5%; *** at 1%. Standard errors (in brackets) are
clustered at industry level for all the regressions.
Age
Size
Num. of unrelated to unrelated replacement
Num. of unrelated to family replacement
Num. of family to unrelated replacement
Num. of family to family replacement
Num. of family manager departures
Num of manager departures
Acquired in two years
(3)
(4)
(5)
Table 10: Merge and Aquisition after managerial team changes
35
0.1677**
(0.0810)
-0.2260***
(0.0568)
0.0150***
(0.0025)
0.0010***
(0.0001)
0.1487**
-0.0606
-0.1651***
(0.0617)
0.0023**
(0.0010)
0.0007***
(0.0001)
-0.0586
(0.0489)
0.0351
(0.0304)
0.0128***
(0.0020)
0.0003***
(0.0001)
0.0708*
(0.0425)
-0.1347***
(0.0516)
0.0006
(0.0008)
0.0004***
(0.0001)
0.0005
(0.0180)
-0.0221
(0.0211)
0.0002
(0.0004)
0.0002***
(0.0000)
0.0482**
(0.0221)
-0.0180
(0.0241)
0.0015***
(0.0005)
0.0001***
(0.0000)
Outcomes of family departures
No rep
Family suc Unrelated suc
(4)
(5)
(6)
-0.0441
(0.0304)
-0.0154
(0.0301)
0.0086***
(0.0013)
0.0002***
(0.0000)
0.0279
(0.0182)
-0.0361**
(0.0171)
0.0031***
(0.0006)
0.0000
(0.0000)
0.0351***
(0.0131)
-0.0094
(0.0091)
0.0011***
(0.0003)
0.0000***
(0.0000)
Outcomes of unrelated departures
No rep
Family suc Unrelated suc
(7)
(8)
(9)
Num. of Obs.
117,051
117,051
117,051
117,051
117,051
117,051
117,051
117,051
117,051
Adjusted R2
0.005
0.002
0.007
0.001
0.001
0.000
0.005
0.002
0.001
Num. of Ind.
177
177
177
177
177
177
177
177
177
Year FE
YES
YES
YES
YES
YES
YES
YES
YES
YES
Industry FE
YES
YES
YES
YES
YES
YES
YES
YES
YES
Sample
Full
Full
Full
Full
Full
Full
Full
Full
Full
Notes: The table shows the reduced form regression results over full sample. The dependent variables are the number of departures
in first three columns, the outcome of family manager departures in the next three columns, and the outcomes of unrelated manager
departures in last three columns. All the dependent variable numbers are calculated in two years forward window. Column (1)
shows the regression results with total number of departures, and column (2) and (3) displays the number of family departures
and unrelated departures respectively. The dependent variables in column (4) and (7) are number of no replacement after family
manager departure and after unrelated manager departure. Similarly, column (5) and (8) shows the results when the dependent
variables are family replacement after family or unrelated managers departures. Column (6) and (9) includes the regression
results with dependent variables of unrelated replacement number after family and unrelated departures respectively. The import
penetration instrumented by current and lagged industry specific exchange indices respectively. Industry exchange rate is weighted
average of exchange rates at the industry level, where the weight is the import share from each country in 2002. The independent
variable import penetration is the value of import divided by the sum of import and the domestic production. Assets is the total
assets in the balance sheet each year in million £. All regressions control for year and industry fixed effects and are estimated at
firm-year level. * significant at 10%; ** at 5%; *** at 1%. Standard errors (in brackets) are clustered at industry level for all the
regressions.
Age
Total Assets
Exch. Rates
Lagged EX
Number of departures
Total
Family
Unrelated
(1)
(2)
(3)
Table 11: Robustness tests: reduced form regressions full sample
36
0.2251***
(0.0709)
-0.2150***
(0.0663)
0.0127***
(0.0020)
0.0013***
(0.0001)
0.1905***
(0.0597)
-0.1705***
(0.0614)
0.0019*
(0.0010)
0.0010***
(0.0001)
-0.0433
(0.0383)
0.0523*
(0.0293)
0.0108***
(0.0015)
0.0003***
(0.0001)
0.1166***
(0.0402)
-0.1476***
(0.0521)
-0.0000
(0.0006)
0.0005***
(0.0001)
-0.0038
(0.0218)
-0.0222
(0.0236)
0.0002
(0.0005)
0.0003***
(0.0000)
0.0526**
(0.0257)
-0.0115
(0.0285)
0.0017***
(0.0006)
0.0002***
(0.0000)
Outcomes of family departures
No rep
Family suc Unrelated suc
(4)
(5)
(6)
-0.0267
(0.0275)
0.0002
(0.0245)
0.0073***
(0.0011)
0.0002***
(0.0000)
0.0244
(0.0207)
-0.0272**
(0.0121)
0.0027***
(0.0006)
0.0001**
(0.0000)
0.0369**
(0.0151)
-0.0175*
(0.0106)
0.0008*
(0.0004)
0.0000**
(0.0000)
Outcomes of unrelated departures
No rep
Family suc Unrelated suc
(7)
(8)
(9)
Num. of Obs.
93,598
93,598
93,598
93,598
93,598
93,598
93,598
93,598
93,598
Adjusted Rˆ2
0.006
0.003
0.007
0.001
0.002
0.001
0.004
0.002
0.001
Num. of Ind.
177
177
177
177
177
177
177
177
177
Year FE
YES
YES
YES
YES
YES
YES
YES
YES
YES
Industry FE
YES
YES
YES
YES
YES
YES
YES
YES
YES
Sample
Founder
Founder
Founder
Founder
Founder
Founder
Founder
Founder
Founder
Notes: The table shows the reduced form regression results over subsample of founder family firms. The dependent variables are
the number of departures in first three columns, the outcome of family manager departures in the next three columns, and the
outcomes of unrelated manager departures in last three columns. All the dependent variable numbers are calculated in two years
forward window. Column (1) shows the regression results with total number of departures, and column (2) and (3) displays the
number of family departures and unrelated departures respectively. The dependent variables in column (4) and (7) are number of
no replacement after family manager departure and after unrelated manager departure. Similarly, column (5) and (8) shows the
results when the dependent variables are family replacement after family or unrelated managers departures. Column (6) and (9)
includes the regression results with dependent variables of unrelated replacement number after family and unrelated departures
respectively. The import penetration instrumented by current and lagged industry specific exchange indices respectively. Industry
exchange rate is weighted average of exchange rates at the industry level, where the weight is the import share from each country
in 2002. The independent variable import penetration is the value of import divided by the sum of import and the domestic
production. Assets is the total assets in the balance sheet each year in million £. All regressions control for year and industry fixed
effects and are estimated at firm-year level. * significant at 10%; ** at 5%; *** at 1%. Standard errors (in brackets) are clustered
at industry level for all the regressions.
Age
Total Assets
Exch. Rates
Lagged EX
Number of departures
Total
Family
Unrelated
(1)
(2)
(3)
Table 12: Robustness tests: reduced form regressions founder family firm sample
37
-0.0855
(0.2194)
-0.2885*
(0.1533)
0.0228***
(0.0080)
-0.0004
(0.0003)
-0.0341
(0.1314)
-0.1463
(0.1342)
0.0032
(0.0021)
-0.0005***
(0.0001)
-0.1433
(0.1585)
-0.0312
(0.0919)
0.0196***
(0.0063)
0.0001
(0.0002)
-0.1313
(0.1090)
-0.0825
(0.1105)
0.0027
(0.0026)
-0.0001
(0.0001)
0.0184
(0.0431)
-0.0206
(0.0360)
-0.0001
(0.0006)
-0.0001*
(0.0001)
0.0287
(0.0541)
-0.0438
(0.0382)
0.0007
(0.0009)
-0.0001
(0.0001)
Outcomes of family departures
No rep Family suc Unrelated suc
(4)
(5)
(6)
-0.1239
(0.1127)
-0.0889
(0.1192)
0.0130***
(0.0047)
0.0002
(0.0002)
0.0429
(0.0387)
-0.0761
(0.0572)
0.0045***
(0.0017)
-0.0001**
(0.0001)
23,453
0.002
153
YES
YES
NF
0.0295
(0.0326)
0.0228
(0.0288)
0.0021*
(0.0011)
0.0001
(0.0000)
Outcomes of unrelated departures
No rep
Family
Unrelated suc
(7)
(8)
(9)
Num. of Obs.
23,453
23,453
23,453
23,453
23,453
23,453
23,453
23,453
Adjusted Rˆ2
0.005
0.001
0.008
0.001
0.001
0.000
0.005
0.003
Num. of Industry
153
153
153
153
153
153
153
153
Year FE
YES
YES
YES
YES
YES
YES
YES
YES
Industry FE
YES
YES
YES
YES
YES
YES
YES
YES
Sample
NF
NF
NF
NF
NF
NF
NF
NF
Notes: The table shows the reduced form regression results over subsample of nonfounder family firms. The dependent variables
are the number of departures in first three columns, the outcome of family manager departures in the next three columns, and the
outcomes of unrelated manager departures in last three columns. All the dependent variable numbers are calculated in two years
forward window. Column (1) shows the regression results with total number of departures, and column (2) and (3) displays the
number of family departures and unrelated departures respectively. The dependent variables in column (4) and (7) are number of
no replacement after family manager departure and after unrelated manager departure. Similarly, column (5) and (8) shows the
results when the dependent variables are family replacement after family or unrelated managers departures. Column (6) and (9)
includes the regression results with dependent variables of unrelated replacement number after family and unrelated departures
respectively. The import penetration instrumented by current and lagged industry specific exchange indices respectively. Industry
exchange rate is weighted average of exchange rates at the industry level, where the weight is the import share from each country
in 2002. The independent variable import penetration is the value of import divided by the sum of import and the domestic
production. Assets is the total assets in the balance sheet each year in million £. All regressions control for year and industry fixed
effects and are estimated at firm-year level. * significant at 10%; ** at 5%; *** at 1%. Standard errors (in brackets) are clustered
at industry level for all the regressions.
Age
Total Assets
Exch. Rates
Lagged EX
Number of departures
Total
Family
Unrelated
(1)
(2)
(3)
Table 13: Robustness tests: reduced form regressions nonfounder family firm sample
38
39
Appendix 1
Variable Definitions
Variable
Dependent Variable
Leave
Transition
Family Succession
Definitions
Leave is a dummy variable defined at firm level. This variable takes the
value of 1 if the firm have at least one manager leave the firm, and takes the
value of 0 otherwise.
Transition is a dummy variable defined at firm level. This variable takes the
value of 1 if the firm have at least one resigned manager matched to a new
coming manager within 1 year gap, 0 otherwise.
Family transition is a dummy variable defined at firm level. This variable
takes the value of 1 if the Transition is 1 and the new manager has the same
family name as the leaving manager.
Competition Measurement
Import Penetration Import penetration is defined as Import value/(Import value+Domestic
production value) at each industry level. Import value comes from Eurostat’s
Comext database and it is measured as the aggregate import (in thousands
EUR) from all partner countries around the world. Domestic production
value is obtained from Eurostat’s Structural Business Statistics database
(SBS) (in thousands EUR) .
Export Open.
Export Openness is defined as Export value / (Export value +Domestic
production value) for each industry. Export value comes from Eurostat’s
Comext database and it is is measured as the aggregate import (EUR
thousands) from all partner countries around the world. Domestic
production value is obtained from Eurostat’s SBS database.
Exch. Rate
The exchange rate is measured at industry level. It is the weighted average
of bilateral exchange rates between UK and its main trading partners. The
weight is the import share of total import value.
Control Variables
Age
Ln Assets
Firm age is measured in years since incorporation.
Natural logarithm of total assets (in million Pounds) .