1

BANK OWNERSHIP STRUCTURE, MARKET DISCIPLINE AND RISK: EVIDENCE FROM A SAMPLE OF PRIVATELY OWNED AND PUBLICLY HELD

EUROPEAN BANKS

Thierno Amadou Barry, Laetitia Lepetit and Amine Tarazi1

Université de Limoges, LAPE, 5 rue Félix Eboué, 87031 Limoges Cedex, France

May 2009 Abstract

The objective of this paper is to analyze the influence of ownership structure on the risk

taking behavior of European commercial banks. We consider five categories of shareholders

(managers/directors, institutional investors, non financial companies, individuals and families,

and banks). Controlling for various factors, we find that asset risk is lower for banks where a

higher proportion of total stocks is held by families and individuals who have less diversified

portfolios. We also find that the probability of default of banks is higher when non financial

companies or institutional investors hold a higher proportion of total equity. However, these

results do not hold for listed banks in which non financial companies hold higher stakes

suggesting that the market might be limiting the risk-taking incentives of such shareholders.

We further show that market forces might be more effective in influencing risk in banks with

a higher involvement of non financial companies than in banks with a higher portion of stock

held by institutional investors. Since listed banks with higher stakes of institutional investors

exhibit higher profitability than their non-listed counterparts, a difference that is not observed

for banks with a higher portion of stock held by non-financial companies, our results suggest

that only inefficient higher risk-taking (not rewarded by higher expected return) is curbed by

market forces.

Keywords: Ownership structure, bank risk, European banks, Market Discipline

JEL Classification: G21, G32 1 Corresponding authors: Tel: +33-555-14-92-05, thierno.barry@unilim.fr (T. Barry); laetitia.lepetit@unilim.fr (L. Lepetit); amine.tarazi@unilim.fr (A. Tarazi).

2

1. Introduction

The last three decades have been characterized by repeated banking crises (the current

financial crisis of 2008, the US savings and loans debacle of the eighties, the 1994-95

Mexican crisis, the 1997 Asian and 1998 Russian financial crises, etc.). Such episodes

highlight the inherently unstable nature of banking and the tendency that banks have towards

excessive risk-taking. In this paper, we aim to focus on one of the driving forces behind the

risk-taking incentives of banks, namely shareholders’ behavior and their incentives to take

higher risk. The issue of ownership structure is of particular interest for the banking industry

as several factors interact and alter governance, such as the quality of bank regulation and

supervision and the opacity of bank assets. Moreover, banking systems faced major changes

during the last 20 years. With financial deregulation and market integration, the scope of

activities of banks has been completely reshaped ranging from traditional intermediation

products to an array of new businesses. These trends led to a substantial consolidation in the

banking industry and consequently to significant changes in ownership and capital structure.

Also, institutional ownership of common stock has increased substantially over the past

twenty years. In terms of shareholding size, expertise in processing information and

monitoring managers, institutional investors (investment companies, investment advisors,

pension funds, etc.) are very different from atomistic individual investors. This might also

imply changes in corporate governance and in banks’ behavior in terms of risk-taking.

However, it is also well known that for publicly traded banks risk-taking incentives

can be mitigated by market forces, and therefore such developments cannot be assessed

without considering incentives driven by financial markets in terms of discipline (Bliss and

Flannery, 2002; Flannery, 2001). In the new Basel Capital Accord, market discipline is one of

three pillars, along with capital regulation (Pillar 1) and banking supervision (Pillar 2). The

idea is to rely on market forces to enhance banking supervision and therefore market

discipline is expected to play an important role. In this context, our goal is to check if market

discipline is actually effective in influencing the risk-taking incentives of different types of

shareholders. To our knowledge there has been no research on whether risk-taking behavior is

different in privately owned banks and publicly owned banks under different ownership

profiles. Kwan (2004), working on a sample of US bank holding companies (BHC), finds that

loan quality and earnings variability are not different between traded BHCs and privately held

BHCs. One of our aims is to assess the risk-taking behavior of banks by combining the two

interrelated dimensions that are ownership structure and market discipline.

3

It has been stressed in the theoretical and empirical literature that agency problems and

risk-taking behavior are different according to the nature of the shareholder. A first issue is

the conflict of interest between managers and shareholders identified by Jensen and Meckling

(1976). Theory indicates that shareholders with a diversified portfolio are motivated to take

more risk for a higher expected return whereas managers take less risk to protect their

position and personal benefits, and preserve their acquired human capital (Galai and Masulis,

1976; Esty, 1998; Jensen and Meckling, 1976; Demsetz and Lehn, 1985). Empirically,

Saunders et al (1990) are the first to test the relationship between banks’ ownership structure

and their risk-taking incentives. They find a positive relationship between managerial stock-

ownership (proportion of stock held by managers) and risk taking. Moreover, they find that

banks controlled by shareholders take more risk than banks controlled by managers. A

number of studies, following Saunders et al. (1990), find a significant effect of ownership

concentration on risk-taking but without any consensus on the sign of such a relationship. If

some studies find a negative relationship, others obtain U-shaped relationships (or inverse U-

shaped) between ownership and risk (Gorton and Rosen, 1995; Chen, et al., 1998; Anderson

and Fraser, 2000). U-shaped relationships between ownership and risk-taking could be

explained by managers’ entrenchment. Moreover, Sullivan and Spong (2007) show that stock

ownership by hired managers is positively linked with bank risk, meaning that under certain

conditions hired managers operate their bank more closely in line with stockholder interests.

Another issue well developed in the literature is the comparison of the performance

(profitability and asset quality) of state-owned banks compared to their private counterparts

(domestic and foreign banks). Agency costs within government bureaucracy can result in

weak managerial incentives and misallocation of resources. According to the agency cost

view, managers exert less effort than their private counterparts or divert resources for personal

benefits, such as, for example, career concerns. For the political view of state ownership,

government-owned banks are inefficient because of the politicians’ deliberate policy of

transferring resources to their supporters (Shleifer, 1998; Shleifer and Vishny, 1986). It has

been underlined that state-owned banks have poorer loan quality and higher default risk than

private-owned banks (Berger et al., 2005; Iannotta et al., 2007). Iannota et al, 2007 also

highlight that mutual banks and government-owned banks appear as less profitable than

private-owned banks. Moreover, they find that government-owned banks have poorer loan

quality and higher default risk, while mutual banks have better loan quality and lower asset

risk than both private and government-owned banks. In addition, some papers have shown

4

that foreign-owned banks exhibit a higher performance than other banks, particularly in

developing countries (Claessens et al., 2001; Bonin et al, 2005; Micco et al., 2007).

Beside the issues of the manager-owner conflict and the differences between state and

private-owned firms, there are other aspects that are well developed in the literature on non

financial firms but not in the literature on financial firms. First, institutional investors who

exercise significant voting power can shape the nature of corporate risk taking. Institutional

investors can exert greater control for reasons of economies of scale in corporate supervision.

Pound (1988) highlights that institutional investors can exercise a control at a lower cost as

they have more experience. There is also the possibility, however, that managers and

institutional investors form an alliance, so that insider interests could take priority over the

maximization of firm value. At the same time, as institutional investors have a diversified

portfolio of investments, they may have fewer incentives to exercise control. Empirical

evidence (Acker and Athanassakos, 2003), based on non financial firms, do not show

conclusive results on the effect of control by institutional investors on firm value. Second,

family-owned firms are perceived as less willing to take risk but also as less profitable. More

generally, firms with large, undiversified owners such as founding families may forgo

maximum profits because they have an undiversified wealth and they are unable to separate

their financial preferences from those of outside shareholders. Families also limit executive

management positions to family members, suggesting a restricted labor pool from which to

obtain qualified and capable talent, potentially leading to competitive disadvantages relatively

to non family-owned firms (Morck et al, 2000). However, James (1999) posits that families

have longer investment horizons, leading to greater investment efficiency. Stein (1988, 1989)

shows that the presence of shareholders with relatively long investment horizons can mitigate

the incentives for myopic investment decisions by managers. Regarding the banking industry,

few papers analyze this issue. Laeven (1999) considers different forms of bank ownership

including state-owned, foreign-owned, company-owned and family-owned banks but not

banks owned by institutional investors. Working on a panel of Asian banks before the Asian

crisis of 1997, he finds that family-owned banks were among the most risky banks together

with company-owned banks whereas foreign-owned banks took little risk relatively to other

banks.

The objective of this paper is to extend the current literature dedicated to the risk-

taking incentives of bank shareholders in several directions. First, we work on a broader

classification of shareholders by considering the equity held by managers, institutional

5

investors, non financial companies, individuals and families, banks, foundations/research

institutes and governments. Second, we consider the proportion of equity held by each

category of owner, instead of using dummy variables to divide ownership into mutually

exclusive categories as in most of the previous studies on bank ownership (Berger et al.,

2005; Bonin et al., 2005; Boubakri et al., 2005; Williams and Nguyen, 2005). This approach

allows us to analyze how the interaction of equity held by different types of shareholders

influences the risk-taking behavior of banks. It also allows us to study the link between

ownership structure and risk more deeply by dealing with the issue of possible coalitions

among different categories or groups. Nevertheless, for consistency with previous studies we

also study the link between risk and the nature of the main shareholder. Third, by

investigating the link between ownership structure and risk for both listed (publicly held) and

non-listed (privately owned) banks we question the ability of market forces to influence bank

risk-taking behavior (market discipline) under different ownership arrangements. Fourth,

previous studies that use a detailed breakdown of the stakes held by different categories of

owners were mostly dedicated to US banks and could not consider as many categories of

shareholders because ownership of banks by non-financial companies is not permitted. By

working on European banks we are therefore able to introduce an additional category which

the literature considers as playing a very controversial role in the management of financial

institutions. Studies on European banks have focused on the nature of ownership (public,

private, mutual, cooperative…) rather than on the structure of ownership in private banks. We

focus on commercial banks only, that is firms that are assumed to have identical objectives,

and to our knowledge this is the first study that looks at the relationship between ownership

structure and risk for European commercial banks.

We work on a panel of European banks through the period 1999-2005. Two main

results emerge from our study. First, we find that banks with different types of ownership

structures have different attitudes in terms of risk-taking. We find a negative relationship

between the proportion of stock held by families and individuals and asset risk, and a positive

relationship between default risk and the proportion of stocks held by institutional investors or

non-financial companies. Second, we find that market forces seem to limit risk-taking

incentives as such results mainly hold for non-listed banks. Nevertheless, we also show that

market forces might be more effective in influencing risk in banks with a higher involvement

of non-financial companies than in banks with a higher proportion of stock held by

institutional investors.

6

The remainder of the paper is structured as follows. Section 2 describes our data and

variables. Section 3 presents the methodology and the hypotheses tested. The empirical results

are discussed in section 4. Section 5 reports robustness checks and discusses further issues.

Section 6 concludes the paper.

2. Data, variables and descriptive statistics

2.1 Data collection and sample definition

The annual data used in this paper are taken from Bankscope Fitch IBCA which

provides information on financial statements and ownership structure for financial institutions

worldwide. We collect the percentage of stocks held respectively by managers and directors,

institutional investors, non-financial companies, self ownership, individual and family

investors, banks, foundations/research institutes, government, unnamed private shareholders

and other unnamed shareholders. Bankscope Fitch IBCA also provides for listed banks the

percentage of stocks held by the public. We use a sample consisting of an unbalanced panel of

annual report data from 1999 to 2005 for a set of European commercial banks established in

16 European countries: Austria, Belgium, Denmark, Finland, France, Germany, Greece,

Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden, Switzerland and United-

Kingdom. We identify in Bankscope 1586 commercial banks for which income statements

and balance sheets are provided for the period 1999-2005. We delete all the banks with less

than five years of time series observations2, which leaves us with 688 banks. Out of this

number of banks, we isolate 320 banks for which detailed data on direct ownership are

available for the years 2001, 2003 and 2005 in the annual financial statement3. Eventually, we

apply other selection criteria and end up with a smaller sample of banks. First, we only

consider banks with a stable ownership structure by comparing the proportion of equity held

by the main shareholders over the period 1999-2005. This restriction is important to

accurately analyze the impact of ownership structure on the performance and risk of banks.

Since our aim is to focus on the influence of different categories of shareholders on

management we need to exclude short run ownership and hit and run strategies that will not

shape the behavior of management and therefore bank risk/profitability in a given direction. 2 This condition enables us to accurately compute the standard deviations of some variables to define risk indicators. 3 Each annual financial statement provides information on the ownership structure of banks for the current year and the previous two years. The report of the year 2001 therefore gives information on the ownership structure of the years 1999, 2000 and 2001. In our study, we consider the direct owner which can be different from the ultimate owner (for example 20% of a bank’s stocks can be owned by a firm (direct owner) in which a family might have a stake of 10%...). Our approach consists in considering the different categories that directly exert control and vote on the bank’s board.

7

We hence only keep banks for which the ownership shares of the main shareholders fluctuate

by less than 10% over the considered period. 249 banks are consistent with this criteria which

enables us to work on a firm-level homogeneous sample. The final sample consists of 249

European commercial banks, within which 80 are listed publicly traded banks4. Among these

banks, 191 banks have a major shareholder with a stake above 50% throughout the whole

sample period and 58 banks (out of which 44 are listed) exhibit ownership shares by the main

shareholders fluctuating by less than 10% (see Table A1 in the appendix for further details on

the distribution of banks by country). We also consider a subsample that satisfies the criteria

that the sum of the different shares that are displayed in Bankscope is at least equal to 99%5.

This criteria leaves us with 198 banks, within which 29 are listed. We test the robustness of

our results by running our estimations on both the large sample of 249 banks and on the

restricted sample of 198 banks.

Descriptive statistics of our large sample of 249 banks are presented in Table 1. We

use data from consolidated accounts if available and from unconsolidated accounts otherwise.

Insert Table 1 here

2.2 Ownership variables

In our study, we code the ownership structure based on the stockholder information

contained in the BankScope database. Two criteria are used to select the categories of owners.

First, we require each category of owner to hold a positive percentage of equity in at least 5

banks. This criteria leads us to exclude three categories of owners, which are Government,

self owned and foundation. Second, we only consider the categories of owners for which we

are able to identify their nature, behavior and incentives to take risk. We therefore exclude

three categories of owners provided by BankScope: public, unnamed private shareholders and

other unnamed shareholders.

Consequently, we end up with five categories of owners that are considered in our

study: (i) managers/directors (MANAGER); (ii) non financial companies (COMPANY); (iii)

individual and family investors (FAMILY); (iv) banks (BANK); and (v) institutional investors -

insurance company, financial companies and mutual & pension funds - (INSTITUT). We

4 Our full dataset contains 137 listed banks. We need to delete: (i) 7 banks with less than five years of time series observations; (ii) 31 banks for which ownership is not detailed in three reports provided for the years 2001, 2003 and 2005; (iii) 19 banks that exhibit a change in ownership structure between 1999 and 2005. 5 The data on ownership structure provided by Bankscope (% share of each type of owner) do not always add up to 100%, particularly for listed banks because we do not always have the percentage held by the public.

8

create five variables which report for each bank in our sample the proportion of equity held by

each category of owner.

Table 2 shows that managers hold equity in only 8 banks out of which 7 are listed

banks. Table 3, which provides statistics on the percentage of equity held by the different

types of owners, also highlights that the proportion of stocks held by managers is very low

(0.30%) compared to the other types of owners. Individuals and families are also involved in

a relatively few number of listed and non-listed banks (25 banks) in our sample of European

commercial banks (see Table 2). Individuals and families are more often involved in listed

banks (see Table 2) but they hold a higher proportion of equity in privately-owned banks

(2.68%) than in listed banks (1.17%) (see Table 3). Institutional investors hold equity in 55

banks. The proportion of stock held by institutional investors is on average equal to 7.81%

(see Table 3) and this category of shareholders is more focused on listed banks (see Table 2).

Non-financial companies are strongly involved in commercial banks as they hold equity in 78

banks out of the 249 banks of our sample. Companies are more often involved in listed banks

but they hold a higher proportion of equity in non-listed banks (see Tables 2 and 3). The

major shareholders of banks are other banking institutions with an average of 58.92% of

equity but mainly in non-listed banks (Table 2). The proportion of equity held by other

banking institutions is higher in non-listed banks (74.88%) than in listed banks (25.23%)

(Table 3).

We compare our sample with the larger population of banks contained in Bankscope

by looking at possible differences between the importance of each category of owner in our

sample of 249 banks and those of the largest sample of 905 banks for which Bankscope Fitch

IBCA provides information on the ownership structure in 2005. The frequencies of banks for

which each category of owner holds a positive percentage of equity in our sample (see Table

2) are not significantly different from those of the largest sample of 905 banks (see Table A2

in appendix). Similarly, the average percentage of equity held by the five categories of owners

that we consider is not significantly different in our sample of 249 banks and in the larger

sample of 905 banks.

Insert Tables 2 and 3 here

Table 4 displays the distribution of the proportion of equity held by the different

categories of owners. The proportion of equity of each category of owner (except managers)

are well distributed in the interval ]0-100]. Our data therefore allow us to consider the

9

proportion of equity held by each category of shareholders to analyze how the involvement of

a given category of shareholders can influence the risk taking behavior of banks.

Insert Table 4

2.3 Risk variables

Table 4 provides statistics for different measures of asset risk and default risk

commonly used in the literature. We compute three standard measures of risk for each bank

throughout the period based on annual accounting data: the standard deviation of the return on

average assets (SD_ROA), the standard deviation of the return on average equity (SD_ROE),

and the mean of the ratio of loan loss provisions to net loans (M_LLP). We also compute

default risk measures. First we use the “Z-score” proposed by Boyd and Graham (1986)

which indicates the probability of failure of a given bank (Z)6. Second, we use the ZP Score

(ZP) as in Goyeau and Tarazi (1992) and Lepetit et al. (2008) and its two additive

components7 (ZP1 and ZP2). ZP1 is a measure of bank portfolio risk whereas ZP2 is a

measure of leverage risk. In table 4, we present the mean of each risk and default indicator for

each of our five categories of owners according to the proportion of equity they hold. Table 4

shows sufficient heterogeneity in different types of shareholders, enabling us to analyze the

behavior of banks depending on their ownership structure.

3. Method and hypotheses tested

Our objective is first to analyze how the proportion of equity held by different types of

shareholders influences the risk-taking behavior of European commercial banks. Second, our

aim is also to investigate whether market discipline can influence the relationship between

ownership structure and risk. We therefore test two hypotheses by specifying two

specifications of our model.

Hypothesis 1: Shareholders’ attitude toward risk is different which implies that banks with

different ownership structures are more or less risky.

We use the following econometric model to test hypothesis 1:

6 )Z (100 average ROE / SDROE= + where ROE and SDROE are expressed in percentage. Higher values of Z-scores imply lower probabilities of failure.

7 ZP=ZP1 + ZP2 = average ROA average (Total equities / Total assets)

SDROA SDROA+ .

10

Model 1

i 0 1 i 2 i 3 i 4 i

5 i 6 i 7 i 8 i i

RISK MANAGER FAMILY INSTITUT COMPANYLISTED M_LNTA M_EQUITY M_DEPOSIT

= α + α +α +α + α ++α +α + α + α + ε

The dependant variable RISKi is a measure of either asset risk or default risk.

Alternatively, we also consider two other dependent variables, which are the mean of the

return on average asset (M_ROA) and the mean of the return on average equity (M_ROE) to

investigate the link between ownership structure and bank profitability.

We consider five categories of owners that may influence the incentives of banks to

take on more risk. The variable MANAGER is the proportion of stock held by

managers/directors. When a manager/director holds a small share of the bank’s equity, she

may have incentives to take less risk. If the bank fails, she loses both her reputation and

human capital investment. This variable is very close to the proxy used by Saunders et al

(1990) which is estimated as the number of shares held by executive and directors divided by

the total of shares outstanding. Note that the underlying assumption in the literature is that a

low proportion of stocks held by managers is associated with a low share of the bank’s stocks

in the managers’ non-human wealth. Also, a higher proportion of stocks held by managers is

assumed to align their interest with those of shareholders as long as the larger investment in

the bank’s stocks does not prevent them from holding diversified portfolios. The relationship

between risk and MANAGER is therefore expected to be positive ( 1 0α > )8 (Saunders et al.,

1990; Knop et Teall, 1996; Anderson and Fraser, 2000) as long as the increase in

managers/directors’ stock holding does not prevent them from holding diversified portfolios.

The variable FAMILY is the proportion of stocks held by individuals and families. We

expect the coefficient of FAMILY to be negative ( 2 0α < ). In general, their portfolio is less

diversified than those of other shareholders. They have incentives to take less risk because if

the bank fails they lose more compared to other shareholders.

The variable INSTITUT is the proportion of stocks held by institutional investors.

Institutional investors will encourage more risky activities that maximize bank value because

their portfolios are sufficiently diversified. The expected coefficient of this variable is positive

( 3 0α > ). However, institutional investors that do not engage in long term investments are less

motivated to control managers. Also, as argued above, since institutional investors hold

diversified portfolios, they might have lower incentives to exert control and therefore the 8 We give here the expected sign for the measures of asset risk (SDROA, SDROE and M_LLP). We expect the opposite sign for the default risk measures (Z and ZP) as the probability of failure is higher when the value of the Z-score is lower.

11

coefficient of this variable might not be significant. In our study, INSTITUT is defined in such

a way that only stable stakes of such investors are taken into account. As explained above, the

observations for which the proportions of stakes are significantly different over time are

omitted.

We also consider shares held by non-financial companies (COMPANY). Banks with a

large portion of stocks held by firms are prone to increase the riskiness of loans granted to

owners. Moreover, if a bank is behind an industrial group, the group management will have

incentives to manipulate the bank to maximize the wealth of ultimate owners. We therefore

expect a positive coefficient associated to the variable Company ( 4 0α > ).

The fifth category of shareholders is banks (BANK). As we can see in Tables 2 and 3, banks

have important stakes in other banks. We expect a negative and significant coefficient for the

variable BANK if banks as shareholders will encourage relatively conservative risk-taking

strategies for reputation concerns. However, we need to remove this variable from Model 1 to

avoid multicolinearity9, particularly for banks for which the sum of the five ownership

components (MANAGER, FAMILY, INSTITUT, COMPANY and BANK) equals 99%10. In this

case, the estimated coefficient associated to each ownership component has to be interpreted

as the effect of a substitution between this component and the BANK component (see

Appendix 2 for details).

A set of control variables are introduced to account for size differences (natural

logarithm of total assets M_LNTA), business differences (deposits to total assets (M_DEP))

and leverage differences (M_EQUITY). Alternative control variables (the ratio of loans to

total assets and the ratio of net non-interest income to net operating income), as well as the

mean of the annual growth rate of total assets to capture the effect on risk of growth strategies

and acquisitions, are also introduced to check for robustness. Because M_LNTA and

M_EQUITY are highly correlated, the leverage ratio is orthogonalized with total assets

(M_OEQUITY). As the information on the ownership structure of our sample of banks is

invariant through time (1999-2005 period) and as our measure of asset risk and default risk

are computed using the standard deviations of ROA and ROE, we conduct cross-section

regressions. We therefore compute the means of our three control variables over the whole

sample period. We also include a dummy variable, LISTED, which takes the value of one if 9 The choice to remove the variable BANK is based on its high correlation both with the variables COMPANY and INSTITUT. 10 We have 198 banks for which the sum of the different percentages that are displayed in Bankscope is at least equal to 99%.

12

the bank is listed on the stock market and zero otherwise. This dummy variable is expected to

capture differences in risk taking for listed and non-listed banks.

To investigate the issue of market discipline, we also test the extent to which market

discipline influences the incentives of different categories of bank shareholders to take risk.

Hypothesis 2: Bank Shareholders’ attitude toward risk depends on the extent of market

discipline. Different categories of shareholders of privately owned (non-listed) banks might

not behave as identical categories of shareholders of publicly owned (listed) banks.

For this purpose, we estimate an augmented model which captures the interaction

between the different categories of owners11 (FAMILY, INSTITUT and COMPANY) and the

dummy variable LISTED which indicates if a bank is listed or not. We therefore use the

following model to test hypothesis 2:

Model 2

i 0 1 i 2 i 3 i 4 i

5 i 6 i 7 i 8 i

9 i i

RISK FAMILY INSTITUT COMPANY FAMILY*LISTEDINSTITUT*LISTED COMPANY*LISTED M_LNTA M_EQUITYM_DEPOSIT

= β + β +β + β + β+β + β +β + β+ β + ε

Interaction variables measure the impact of market discipline on bank shareholders’

attitude toward risk. For example, a positive and significant coefficient associated to

INSTITUT ( 2 0β > ) will indicate that a higher involvement of institutional investors in total

equity of non-listed banks will increase risk. But a negative and significant value of the sum

of the coefficients of the variable INSTITUT and the interaction variable

INSTITUT*LISTED ( 2 5 0β +β < ) will imply the opposite result for listed banks. If the sum

of these two coefficients is not significantly different from zero, then our model will highlight

the absence of any link between the proportion of stocks held by such investors and risk for

listed banks. As for Model 1, we remove the variable BANK in our estimations to avoid

multicolinearity. Again, the estimated coefficient associated to each interaction variable has to

be interpreted as a substitution between each ownership component and the BANK component

(see Appendix 2 for details).

11 As managers hold stocks in only one non-listed bank, we cannot consider the variable MANAGER in Model 2.

13

4. Results

Tables 5 and 6 show the results obtained for Models 1 and 2 with cross section OLS

estimations with t-statistics corrected for heteroskedasticity following White’s methodology. .

Our results are consistent with hypothesis 1. We find that the portion of total equity held by

different categories of shareholders is significant in explaining risk differences.

First, as expected, our results show that there is a negative and significant relationship

between the FAMILY ownership component and the two measures of asset risk (SDROA and

SDROE) (Table 5). Higher portions of total stock held by individuals and families

(compensated in our model by a decrease in the BANK component, (see appendix 2)), are

associated with lower asset risk. As argued above, such shareholders hold less diversified

portfolios and are often involved in the management of such banks. However, our results

highlight that the level of involvement of individuals and families does not influence the

proxy we use for credit risk (M_LLP) as well as bank default risk (Z and ZP).

Second, we find a positive and significant relationship between the proportion of

equity held by institutional investors and the proxy for credit risk M_LLP (Table 5). This

result indicates that loans are more risky in banks where a higher portion of total stocks is

held by institutional investors. Such investors are expected to pursue firm value maximization

strategies and their portfolio is generally well diversified. Interestingly, we also find a

negative and significant coefficient associated to INSTITUT for our different measures of

default risk. The probability of default of banks increases when institutional investors hold a

higher proportion of total equity.

Third, the portion of equity held by non financial companies does not significantly

influence the riskiness of bank assets. This result is not consistent with the findings of

Boubakri et al. (2005) who show that industrial groups-controlled banks are the ones with the

highest risk exposure in developing countries. However, our results show that default risk

(ZP, ZP1 and ZP2) is higher when the portion of shares held by such firms increases.

Finally, we do not find any significant relationship between the level of manager’s

equity and our risk and default measures. This result is not consistent with previous studies on

US banks which find a significant effect of managerial ownership on risk-taking but without

any consensus on the sign of this relationship (Saunders et al,1990; Gorton and Rosen, 1995;

Knopf and Teall, 1996; Chen et al., 1998; Anderson and Fraser, 2000). However, it should be

noted that our data do not allow us to infer any accurate relationship between manager

involvement and risk. As shown by table 2 and table A2, managers rarely hold stocks in their

14

own company in our sample. Moreover, 7 out the 8 banks in which they have a stake are

listed banks.

Insert Table 5 here

Regarding the influence of market forces on risk-taking, the coefficient associated to

the variable LISTED in Model 1 (Table 5) is not significant, except for SDROA and ZP1. At

first sight, there seems to be no significant difference in risk between listed and non listed

banks suggesting that market forces might not strongly influence the risk behavior of listed

banks in a specific way. However, our results show that listed banks exhibit lower income

variability (standard deviation of the ROA) and are more profitable than non-listed banks.

We further investigate the issue of market discipline with Model 2 by considering the

interaction between the portion of equity held by each category of owner and the exposure of

banks to market forces (Table 6)12. We find a negative and significant relationship between

the proportion of stocks held by non-financial companies and the measures of default (ZP and

ZP1) for non-listed banks but not for listed banks. Therefore, our results indicate that the

higher probability of default associated to banks with a higher portion of equity held by non

financial companies only holds for non-listed banks. Market forces might therefore actually

discipline the risk-taking of such banks when they are listed.

Also, a higher default risk is associated to non-listed banks with a larger involvement

of institutional investors. For most of our default risk measures (Z, ZP and ZP2) such a result

also holds for listed banks suggesting that the market might be less effective in influencing

the risk behavior of such banks. However, the coefficient of our proxy of credit risk M_LLP is

no longer significant for listed banks implying that loans are less risky when such banks are

listed and therefore exposed to market forces. Regarding family ownership, we find a

negative and significant relationship between FAMILY and asset risk (standard deviation

ROA) for both non-listed and listed banks. Therefore, our above result showing a lower risk

in banks with a higher stake of family-type shareholders holds for both listed and non-listed

banks.

Insert Table 6 here

12 We do not include MANAGER in Model 2 because, in our sample, only one bank involving managerial shareholding is not listed (see Table2).

15

5. Robustness checks and further issues13

Several robustness checks are performed. We run separate regressions introducing our

ownership variables one by one along with the control variables. All conclusions concerning

the variables of interest remain unchanged. In addition to our previous results, we find, as

expected, that there is a significant and negative relationship between the proxy of credit risk

(M_LLP) and the proportion of total equity held by banks. The results also show that a higher

proportion of equity held by banks is associated to a lower probability of default.

We also estimate Model 1 and Model 2 using the restricted sample of 198 banks for

which the sum of the different equity components is at least equal to 99%. We consider this

restricted sample to ensure that our results are not biased by the fact that some information

regarding ownership structure might be missing or not reported in the Bankscope dataset that

we use. Our conclusions regarding the inclusion of ownership variables remain unchanged.

We further perform a number of robustness checks that are specification related. First,

we include country dummies to capture the presence of country specific effects. Second, other

control variables to account for business differences are introduced in the estimations such as

the ratio of loans to total assets and the ratio of net non-interest income to net operating

income. We also run our estimations by introducing the mean of the annual growth rate of

total assets. An increase in a bank’s total assets is presumed to capture the effect on risk of

growth strategies and acquisitions experienced by many European banks in the early 2000’s.

Our conclusions regarding our ownership variables are unaltered.

Eventually, to further examine issues related to the influence of ownership structure on

the risk-taking behaviour of banks, we carry out a deeper investigation of our sample.

We conduct our estimations separately for large banks (total assets > 1 billion Euros)

and small banks (total assets < 1 billion Euros) to further check for size effects on the

relationship between ownership structure and banks’ behaviour in terms of risk taking. Table

A3 in appendix presents the results obtained for Model 114. Ownership variables are

significant in explaining risk differences for both the sample of large and small banks which

is consistent with hypothesis 1.

13 The results from the estimations conducted in this section are available from the authors on request. 14 The distribution of the proportion of equity held by the different categories of owners for the sample of small banks does not allow us to run our estimations for Model 2.

16

To be consistent with previous studies on ownership in banking, we also classify

banks according to the nature of their main category of shareholders. The objective of such a

classification is to analyze if the risk-taking behavior of banks is different according to the

nature of the main category of shareholders. We consider that there is a majority ownership

when a category of owner holds strictly above 33% of total equity. We have in our sample of

249 banks: (i) 1 manager-owned bank; (ii) 6 family and individual-owned banks15; (iii) 20

institutional investor-owned banks; (iv) 32 company-owned banks, and (v) 156 bank-owned

banks (see Table 4). We create the following four dummy variables which take the value of

one when ownership is higher than 33% of the equity and 0 otherwise: FAMILY_OWNED,

INSITUT_OWNED, COMPANY_OWNED and BANK_OWNED. We do not consider in our

estimations manager-owned banks because only one bank has a majority of equity held by

managers (see Table 4). We also exclude the variable BANK_OWNED due to its high

correlation with the variables INSTITUT_OWNED and COMPANY_OWNED. Table A4 in

appendix shows the results of the estimations for Model 116. We find that the probability of

default is higher when the major category of owners is institutional investors compared to

banks for which the main category of owners is banking institutions. But we do not find any

difference in risk-taking behavior between banks which are majority-owned by non financial

companies and those which are majority-owned by banking institutions. Our results also show

that banks which are majority-owned by families and individuals exhibit a lower credit risk

level. These results are consistent with the hypothesis that the nature of the majority

shareholder influences the risk-taking behavior of banks.

Our results therefore highlight that both the degree of involvement of shareholders and

the nature of the main category of shareholders influence the attitude of banks toward risk.

6. Summary and concluding remarks

The objective of this study was to analyze if the risk-taking behavior of banks is

influenced by their ownership structure. We differentiate five categories of shareholders who

have different risk-taking incentives (managers/directors, institutional investors, non-financial

companies, individual and family investors, and banks). Our aim was also to assess if market

discipline influences the incentives of different categories of bank shareholders to take risk.

Working on a panel of European commercial banks and using both asset risk and

default risk measures, we find that ownership structure is significant in explaining risk

15 Out of these 6 family-owned banks, 2 banks also have one other major shareholder. 16 The number of banks for which we have a majority owner do not allow us to run estimations for Model 2.

17

differences. Specifically, we find that asset risk is lower in banks where a higher proportion of

total stocks is held by families and individuals and that the probability of default of banks is

higher when non-financial companies and institutional investors hold a higher proportion of

total equity. Our results further highlight that market forces might actually discipline the risk-

taking behavior of such investors as the positive relationship between the proportion of equity

held by non-financial companies and default risk only holds for non-listed banks. Our

findings regarding market discipline are however less robust for institutional investors than

for non financial companies. Therefore, our study suggests that market discipline might be

less effective to curb the risk-taking behavior of institutional investors than the shareholder

risk-taking incentives of non-financial companies. A closer look shows that listed banks with

higher stakes of institutional investors exhibit higher profitability than their non-listed

counterparts. Conversely, for banks with a higher involvement of non-financial firms there is

no significant difference in profitability between listed and non-listed institutions. Therefore,

whereas higher default risk is offset by higher profitability for banks dominated by

institutional investors, such a result cannot be observed for banks influenced by non-financial

firms. On the whole, the market might be counteracting the behavior of banks controlled by

non-financial firms by limiting inefficient risk-taking (bad risk) but not the efficient risk-

taking (good risk) of institutional investors.

18

Table 1. Descriptive statistics for our panel of 249 European Commercial banks, on average over the period 1999-2005 LOANS DEP EQUITY EXPENSES LLP ROA ROE LIQUID OBS TA Z SDROA Whole sample (249banks)

Mean 50,15 39,06 9,41 1,58 0,33 0,81 9,00 24,50 32,15 20 200 44,03 0,58Maximum 94,71 93,31 68,24 41,78 9,09 16,59 30,82 87,09 887,90 839 000 511,66 7,67Minimum 0,76 0,00 1,06 0,04 -49,66 -4,04 -119,30 0,24 0,02 4 554 0,56 0,01Std. Dev. 24,66 26,04 8,51 2,80 3,77 1,43 12,30 20,65 73,68 83 900 56,13 0,94Non-listed banks (169)

Mean 45,55 31,65 9,71 1,35 0,18 0,63 7,27 28,36 33,56 3 820 43,63 0,61Maximum 94,71 93,31 66,78 7,73 9,09 4,87 30,82 87,09 887,90 52 400 511,66 7,67Minimum 0,76 0,00 1,47 0,04 -49,66 -4,04 -119,30 0,31 0,02 4 554 0,56 0,01Std. Dev. 25,96 25,98 8,81 1,06 4,56 1,04 13,82 23,18 86,32 7 990 58,64 0,97Listed banks (80)

Mean 59,87 54,71 8,79 2,05 0,66 1,18 12,67 16,33 29,10 54 800 44,87 0,51Maximum 88,84 86,94 68,24 41,78 3,39 16,59 25,55 47,28 141,52 839 000 396,34 5,58Minimum 9,09 3,84 1,06 0,04 -0,98 -2,86 -20,21 0,24 0,90 57 462 1,57 0,01Std. Dev. 18,28 18,09 7,84 4,68 0,54 1,98 7,00 9,87 32,76 142 000 50,83 0,88Variable definitions (all variables are expressed in percentage except TA which is in million of euros): LOANS = net loans/total assets; DEP = deposits/total assets; EQUITY= equity/total assets; EXPENSES =personnel expenses/total assets; LLP = loan loss provision/net loans; ROA = return on average assets; ROE= return on average equity; LIQUID = liquid assets/total assets; OBS= off balance sheet/ total assets; TA= total assets (thousand Euros); SDROA= standard deviation of the ROA; Z = Z-score.

19

Table 2. Number of banks for which the ownership variables are different from zero

MANAGERa FAMILY INSTITUT COMPANY BANK

=0 (%)a

>0 (%)

=0 (%)

>0 (%)

=0 (%)

>0 (%)

=0 (%)

>0 (%)

=0 (%)

>0 (%)

Whole sample (249 banks)

241 (96.78)

8 (3.22)

224 (89.95)

25 (10.05)

194 (77.91)

55 (22.09)

171 (68.67)

78 (31.33)

69 (27.71)

180 (72.29)

Non-listed banks (169 banks)

168 (99.5)

1 (0.5)

160 (94.68)

9 (5.32)

151 (89.35)

18 (10.65)

135 (79.89)

34 (20.11)

34 (79.89)

135 (23.74)

Listed banks (80 banks)

73 (91.25)

7 (8.75)

64 (80.00)

16 (20.00)

43 (53.75)

37 (46.25)

36 (45.00)

44 (55.00)

35 (43.75)

45 (56.25)

The variables MANAGER, FAMILY, INSTITUT, COMPANY and BANK represent the percentage of stock held respectively by managers, families and individuals, institutional investors, non-financial companies, and banks. a For example, in the whole sample, we have 241 banks in which managers do not hold equity and 8 banks in which managers hold a positive percentage of equity. We also present the percentage of banks for which the variable MANAGER is equal to zero (96.78). Table 3. Descriptive statistics of the ownership variables for our panel of 249 European banks (% of stock held by the different categories of owners)

MANAGER FAMILY INSTITUT COMPANY BANK Whole sample (249) Mean 0,30 2,19 7,81 12,36 58,92 Maximum 33,72 100 100 100 100 Minimum 0 0 0 0 0 Std. Dev. 2,51 11,76 22,98 28,15 44,92 Non-listed banks (169) Mean 0,10 2,68 7,49 13,54 74,88 Maximum 16,48 100 100 100 100 Minimum 0 0 0 0 0 Std. Dev. 1,27 14,12 25,12 31,72 40,93 Listed banks (80) Mean 0,74 1,17 7,27 11,13 25,23 Maximum 33,72 17,14 79,86 99,97 99,9 Minimum 0 0 0 0 0 Std. Dev. 4,02 3,00 14,45 20,48 32,84 The variables MANAGER, FAMILY, INSTITUT, COMPANY and BANK represent the percentage of stock held respectively by managers, families and individuals, institutional investors, non-financial companies, and banks.

20

Table 4. Risk measures and default risk measures according to the shareholder type and the percentage of equity held Percentage of equity held 0 ]0-5] ]5-33] ]33-50] ]50-75] ]75-100]

SDROA 0,57 0,47 0,99 0,19 - - SDROE 6,88 9,20 12,23 2,70 - - MLLP_NL 0,32 0,71 0,48 0,57 - - Z 44,60 33,15 18,11 44,91 - - ZP 50,38 33,32 19,14 50,96 - - ZP1 4,25 3,37 2,27 8,93 - - ZP2 46,13 29,95 16,87 42,04 - -

MANAGER

Observation 241 3 4 1 SDROA 0,58 0,57 0,54 0,46 0,12 0,15 SDROE 7,24 5,31 4,51 7,27 2,23 2,27 MLLP_NL 0,34 -0,22 0,71 0,26 0,75 0,14 Z 44,19 28,35 39,21 109,51 49,37 49,96 ZP 50,80 31,09 31,79 95,20 53,26 48,98 ZP1 4,26 3,58 3,55 7,69 4,69 3,22 ZP2 46,54 27,70 28,24 87,51 48,57 45,75

FAMILY

Observation 224 9 10 2 1 3 SDROA 0,57 0,20 0,49 0,34 0,25 1,41 SDROE 7,20 3,11 6,29 4,24 3,56 11,45 MLLP_NL 0,17 0,29 0,74 0,46 1,38 1,80 Z 43,37 86,80 30,75 36,54 65,11 16,43 ZP 49,32 109,34 26,90 43,89 59,90 16,13 ZP1 4,15 8,63 3,29 3,27 4,83 1,49 ZP2 45,16 100,00 23,61 40,61 55,06 14,63

INSTITUT

Observation 194 15 20 3 5 12 SDROA 0,62 0,23 0,30 0,27 0,20 0,93 SDROE 7,19 3,02 4,66 1,83 2,54 14,53 MLLP_NL 0,28 0,49 0,34 0,46 0,64 0,50 Z 37,51 84,66 54,60 107,90 70,97 28,31 ZP 47,84 48,98 67,88 63,66 74,89 24,54 ZP1 3,91 5,60 6,47 5,08 4,69 2,16 ZP2 43,98 43,38 61,41 58,58 70,19 22,38

COMPANY

Observation 171 15 31 4 8 20 SDROA 0,75 0,30 0,37 0,14 0,31 0,58 SDROE 8,78 3,57 6,03 1,79 3,68 7,04 MLLP_NL 0,84 0,52 0,71 1,05 0,83 -0,11 Z 32,15 35,58 47,26 78,66 43,10 48,09 ZP 34,69 34,30 65,79 67,14 46,15 55,22 ZP1 3,72 4,59 7,13 6,50 3,77 3,97 ZP2 30,96 29,71 58,65 60,63 42,37 51,24

BANK

Observation 69 5 19 8 17 131 The variables MANAGER, FAMILY, INSTITUT, COMPANY and BANK represent the percentage of stock held respectively by managers, families and individuals, institutional investors, non-financial companies, and banks. Variable definitions (all the variables are expressed in percentage; standard deviations and means are computed over the period 1999-2005): SDROA= standard deviation of the return on average assets; SDROE = standard deviation of return on average equity, MLLP_NL = Mean of the ratio of loan loss provision to net loans over the sample period; Z = Z-score; ZP = ZP-score; ZP1=measure of bank portfolio risk; ZP2 = measure of leverage risk.

21

Table 5. Influence of ownership structure on the risk-taking behavior and profitability of banks (Model 1), cross-section OLS regressions Risk measures Default Risk measures Profitability measures SDROA SDROE M_LLP Z ZP ZP1 ZP2 M_ROA M_ROE

CONSTANT

0.8666*** (3.525)

8.5219*** (3.278)

0.6652 (0.925)

6.6822*** (5.175)

81.069*** (4.271)

6.9279*** (2.936)

7.141*** (4.337)

0.6529** (2.015)

4.3370* (1.706)

MANAGER

0.0087 (0.689)

0.1148 (0.662)

-0.0074 (-0.573)

-0.6197 (-0.915)

- 0.5596 (-0.761)

-0.0007 (0.008)

-0.5597 (-0.863)

0.0189** (2.406)

0.2594***

(2.687)

FAMILY

-0.0041** (-2.457)

-0.0654*** (-2.961)

-0.0035 (-1.232)

0.1892 (1.060)

-0.0150 (-0.082)

-0.0181 (-1.076)

0.0031 (0.018)

0.0006 (0.274)

0.0043 (0.182)

INSTITUT

0.0034 (1.390)

0.0359 (1.438)

0.0165* (1.929)

-0.2326** (-2.185)

-0.3618*** (-3.053)

-0.0378*** (-2.745)

-0.3240*** (-3.021)

0.0040 (0.915)

0.0022 (0.078)

COMPANY

-0.0005 (-0.287)

0.0578 (0.870)

0.0048 (0.983)

-0.0260 (-0.274)

- 0.1901* (-1.746)

-0.0250** (-2.024)

-0.1651* (-1.650)

-0.0033 (-0.949)

-0.0692 (-1.080)

M_LNTA

-0.0189 (-1.519)

-0.0698 (-0.416)

-0.0044 (-0.062)

-2.2110** (-2.119)

-2.0300 (-1.484)

-0.0700 (-0.493)

-1.9599 (1.528)

0.0169 (0.976)

0.5114***

(3.072)

M_OEQUITY

0.0540*** (7.387)

-0.0954 (-1.400)

-0.0482** (-2.502)

0.0536 (0.180)

-0.3418 (-1.053)

-0.0623 (-1.424)

-0.2794 (-0.962)

0.0997** (3.156)

0.0241 (0.308)

M_DEPOSIT

-0.0024 (-0.987)

-0.0287 (-0.886)

-0.0203 (-1.368)

0.0323 (0.189)

-0.1089 (-0.483)

-0.0637*** (-2.763)

-0.0451 (-0.213)

-0.0060* (-1.648)

-0.0453 (-1.640)

LISTED

- 0.1582** (-2.111)

-1.7951 (-1.536)

1.0511 (1.335)

-2.3716 (-0.260)

-5.7991 (-0.449)

3.5788** (2.350)

-9.3780 (-0.787)

0.5580***

(3.410)

(6.8806***

(5.594) Number of obs. 249 249 247 249 249 249 249 249 249 R² 0.2956 0.0367 0.0421 0.0302 0.0222 0.0608 0.0225 0.4257 0.0967 .***, ** and * indicate significance, respectively, at the 1%, 5% and 10%levels respectively. t-statistics are corrected for heteroskedasticity following White’s methodology. Variable definitions (standard deviations and means are computed over the period 1999-2005): SDROA= standard deviation of the return on average assets; SDROE = standard deviation of the return on average equity , M_LLP = mean of the ratio of loan loss provisions to net loans; Z = Z-score; ZP = ZP-score; ZP1=measure of bank portfolio risk; ZP2 = measure of leverage risk; M_ROA= mean of the return on average asset; M_ROE= mean of the return on average equity;; M_LNTA= mean of logarithm of total asset; M_OEQUITY = mean of the ratio of equity to total assets orthogonalized with TA; M_DEPOSIT = mean of the ratio of deposits to total assets; LISTED= dummy variable equal to 1 if the bank is listed on a stock exchange, and 0 otherwise. The variables MANAGER, FAMILY, INSTITUT and COMPANY represent the percentage of stock held respectively by managers, families and individuals, institutional investors and non-financial companies.

22

Table 6. Influence of ownership structure on the risk-taking behavior and profitability of listed and non-listed banks (Model 2), cross-section OLS regressions Risk measures Default Risk measures Profitability measures SDROA SDROE M_LLP Z ZP ZP1 ZP2 M_ROA M_ROE CONSTANT

0.8292***(3.410)

7.9518** (2.436)

0.8681 (1.398)

6.7057*** (5.581)

80.478*** (4.358)

7.6717*** (3.327)

72.806*** (4.386)

0.6953** (2.310)

5.6094** (2.207)

FAMILY

-0.0034* (-1.913)

-0.0608** (-2.620)

-0.0050 (-1.619)

0.2060 (1.124)

0.0092 (0.049)

-0.0176 (-0.999)

0.0268 (0.156)

0.0016 (0.756)

0.0051 (0.182)

INSTITUT

0.0032 (1.255)

0.0338 (1.278)

0.0166* (1.749)

-0.1799* (-1.687)

-0.3255*** (-2.672)

-0.0368*** (-2.706)

-0.2886*** (-2.616)

-0.0012 (-0.465)

-0.0112 (-0.358)

COMPANY

FAMILY*LISTED

-0.0002 (-0.112) -0.0300*

0.0782 (0.367) -0.1454

0.0036 (0.686) 0.0767

-0.0279 (-0.266) -0.8041

-0.1987* (-1.665) -1.2691

-0.0303** (-2.303) 0.0265

-0.1684 (-1.533) -1.2957

-0.0023 (-0.692) -0.0647

-0.0854 (-1.095) 0.1196

INSTITUT *LISTED

(1.748) 0.0020

(-0.790) 0.0293

(1.254) 0.0011

(-0.645) -0.4120**

(-1.191) -0.2787

(0.251) 0.0160

(-1.296) -0.294

(-1.256) 0.0539**

(0.628) 0.1508***

COMPANY*LISTED

(0.412) -0.0017

(0.480) -0.1065

(0.076) 0.0081

(-1.961) -0.0388

(-1.387) 0.0027

(0.750) 0.0376*

(-1.601) -0.0349

(2.040) 0.0013

(2.728) 0.1171

M_LNTA

(-0.494) -0.0169

(-1.300) -.0495

(0.965) -0.0194

(-0.171) -2.2509**

(0.015) -2.0046

(1.782) -0.1322

(-0.209) -1.8724

(0.278) 0.0083

(1.458) 0.4165**

M_OEQUITY

(-1.379) 0.0530**

(-0.300) -0.1160

(-0.303) -0.0438**

(-2.304) 0.1366

(-1.525) -0.2929

(-1.021) -0.0500

(1.508) -0.2428

(0.514) 0.0908***

(2.501) 0.0250

M_DEPOSIT

(7.029) -0.0030

(-1.496) -0.0296

(-2.161) -0.0142

(0.430) 0.0368

(-0.880) -0.1301

(-1.165) -0.0412*

(-0.807) -0.0888

(4.409) -0.0021

(0.295) -0.0087

(-1.049)

(-0.826)

(-1.314)

(0.224)

(-0.646)

(-1.835)

(-0.479)

(-0.677)

(-0.295)

Risk level to reject 1 4 0β + β = 4.64%** 25.36% 23.10% 62.35% 23.18% 93.16% 20.01%

21.92%

50.44%

Risk level to reject 2 5 0β + β = 27.70% 29.76% 10.73% 0.4%*** 0.7%*** 38.85% 0.4%***

4.78%**

0.73%***

Risk level to reject 3 6 0β + β = 52.25% 28.96% 17.27% 74.31% 25.36% 71.92% 19.66%

81.71%

25.97% Number of obs. 249 249 247 249 249 249 249 249 249 R² 0.2938 0.0431 0.0307 0.0398 0.0222 0.0301 0.0217 0.4834 0.0617 .***, ** and * indicate significance, respectively, at the 1%, 5% and 10%levels respectively. t-statistics are corrected for heteroskedasticity following White’s methodology. Variable definitions (standard deviations and means are computed over the period 1999-2005): SDROA= standard deviation of the return on average assets; SDROE = standard deviation of the return on average equity , M_LLP = mean of the ratio of loan loss provisions to net loans; Z = Z-score; ZP = ZP-score; ZP1=measure of bank portfolio risk; ZP2 = measure of leverage risk; M_ROA= mean of the return on average asset; M_ROE= mean of the return on average equity; M_LNTA= mean of logarithm of total asset; M_OEQUITY = mean of the ratio of equity to total assets orthogonalized with TA; M_DEPOSIT = mean of the ratio of deposits to total assets. The variables FAMILY, INSTITUT and COMPANY represent the percentage of stock held respectively by families and individuals, institutional investors and non-financial companies.

23

References Acker L., Athanassakos G., 2003. A simultaneous equation analysis of analyst' forecast bias

and institutional ownership. Journal of Business Finance and Accounting, 30, 1017-1041

Anderson,R.C and. Fraser, D.R, 2000. Corporate Control, Bank Risk Taking, and the Health

of the Banking Industry. Journal of Banking and finance, 24,1383-1398.

Berger, A.N., Clarke, G.R.G., Cull, R., Klapper, Leora., Udell, G.F., 2005. Corporate

governace and bank performance: A joint analysis of the static, selection, and dynamic effects

of domestic, foreign, and state ownership. Journal of Banking and Finance, 29, 2179-2221.

Bliss, R., Flannery, M., 2002. Market discipline in the governance of U.S. Bank Holding

Companies : monitoring vs. influencing. European Finance Review 6 (3), 361-395.

Bonin, J.P., Hasan , I., Wachtel, P., 2005. Bank performance, efficiency and ownership in

transition countries. Journal of Banking and Finance, 29, 31-54.

Boubakri, N., Cosset, J.C., Fischer K., Guedhami O., 2005. Privatization and Bank

Performance in developing countries. Journal of Banking and Finance, 29, 2015-2041.

Boyd J.H., Graham S.L, 1986. Risk, Regulation and Bank Holding Company Expansion.

Quaterly Review, Federal Reserve Bank of Mineapolis.

Chen,C.R.,Steiner,T.L, Whyte, A.M, 1998. Risk-taking Behavior and Management

Ownership in Depository Institutions. Journal of Financial Research, 21,1-16.

Classens, S., Demirguc-kunt, A., Huizinga, H., 2001. How does foreign Entry affect the

domestic banking market ? Journal of Banking and Finance, 25, 891-911.

Demsetz, H., Lehn, K, 1985. The Structure of Corporate Ownership: Causes and

Consequences. Journal of Political Economics, 93, 1155-1177.

Esty,B., 1998.The impact of Contingent Liability on Commercial Bank Risk Taking. Journal

of Financial Economics, 47, 189-218.

Flannery M. J., 2001. The faces of market discipline, Journal of Financial Services Research,

vol. 20, 2-3, 107-119.

Galai, D., Masulis, R., 1976. The option Pricing Model and the Risk Factor of Stock. Journal

of Financial Economics, 3 ,53-81.

Gorton, G., Rosen,R.,1995. Corporate Control, Portfolio Choice, and the Decline of Baking.

Journal of Finance, 50, 509-527.

Goyau, D., Tarazi A., 1992. Evaluation du risqué de défaillance bancaire en Europe. Revue

d’Economie Politique, 102, 249-280.

Iannota, G., Giacomo, N., Sironi, A., 2007. Ownership Structure, Risk and Performance in the

European Banking Industry. Journal of Banking and Finance, 31, 2127-2149.

24

James, H., 1999. Owner as manager, extended horizon and the family firm. International

Journal of the Economics of Business, 6, 41-56

Jensen, M.C., Meckling, W.H.,1976. Theory of the firm: Managerial behavior, Agency cost

and Ownership structure. Journal of Financial Economics, 3, 305-360.

Laeven, L., 1999. Risk and efficiency in East Asian banks. World Bank Discussion Paper

n°2255.

Lepetit, L., Nys, E., Rous, P., Tarazi, A., 2008. Bank Income Structure and Risk: An

Empirical Analysis of European Banks, Journal of Banking and Finance, 32, 1452-1467.

Knopf, J.K.,and Teal,J.,1996.Risk taking behavior in the U.S thrift industry: ownership

structure and regulatory changes. Journal of Banking and Finance, 1, 3-11.

Kwan, HK., 2004. Risk and Return of Publicly held versus privately Owned banks. FRBNY

Economic Policy Review.

Laeven, L., 1999. Risk and efficiency in East Asian banks. World Bank Discussion Paper.

Micco,A., Ugo, P., Monica, Y., 2007. Bank ownership and performance. Does politics

matter? Journal of Banking and Finance, 31, 219-241.

Morck, R., Stangeland, D., Yeung, B., 2000. Inherited wealth, corporate control, and

economic growth: The Canadian disease’, in: Morck, R. (Ed.), Concentrated Corporate

Ownership, National Bureau of Economic Research Conference Volume, University of

Chicago Press.

Pound J., 1988. Proxy contests and the efficiency of shareholder oversight. Journal of

Financial Economics, 10, 53-83

Saunders,A., Strock,E.,Travlos,N,1990. Ownership Structure, Deregulation, and Bank Risk

Taking. Journal of Finance, 13, 2.

Shleifer, A., Vishny, R., 1986. Large Shareholders and Corporate Control. Journal of

Political Economy, 94, 461-488.

Stein, JC., 1989. Efficient capital markets, inefficient firms: A model of myopic corporate

behavior. Quarterly journal of Economics, 655-669.

Stein, JC., 1988 .Takeover threats and managerial myopia. Journal of Political Economy, 61-

68.

Sullivan,R., and Spong, K.R,2007. Managerial Wealth, Ownership Structure, and Risk in

Commercial Banks. Journal of Financial Intermediation, 16, 229-248.

Williams, J., Nguyen, N., 2005. Financial liberalisation, crisis, and restructuring: A

comparative study of bank performance and bank governance in South East Asia. Journal of

Banking and Finance, 29, 2119-2154.

25

Appendix 1

Tableau A1. Distribution of banks by country All banks Non-listed banks Listed banks Austria 14 11 3

Belgium 7 7 0

Denmark 19 2 17

Finlande 2 0 2

France 64 58 6

Germany 39 33 6

Greece 7 0 7

Ireland 5 1 4

Italy 17 4 13

Luxembourg 33 32 1

Netherlands 7 6 1

Portugal 2 0 2

Spain 15 3 12

Sweden 2 0 2

Switzerland 3 2 1

United Kingdom 13 10 3

Total 249 169 80 Table A2. Descriptive statistics of the ownership variables for the large sample of 905 banks for which Bankscope Fitch IBCA provide information on the ownership structure in 2005

MANAGERa FAMILY INSTITUT COMPANY BANK

=0 (%)a

>0 (%)

=0 (%)

>0 (%)

=0 (%)

>0 (%)

=0 (%)

>0 (%)

=0 (%)

>0 (%)

Number of banks

891 (98.46)

14 (1.54)

845 (93.38)

60 (6.62)

724 (80)

181 (20)

649 (71.72)

256 (28.28)

289 (31.94)

616 (68.06)

Percentage of equity 0 0.192 0 2.769 0 9.664 0 18.097 0 62.316

The variables MANAGER, FAMILY, INSTITUT, COMPANY and BANK represent the percentage of stock held respectively by managers, families and individuals, institutional investors, non-financial companies, and banks. a For example, in the whole sample, we have 891 banks for which the managers do not hold equity and 14 banks for which the managers hold a positive percentage of equity. We also present the percentage of banks for which the variable MANAGER is equal to zero (98.46).

26

Table A3. Influence of ownership structure on the risk-taking behavior and profitability of large banks and small banks (Model 1), cross-section OLS regressions Risk measures Default Risk measures Profitability measures SDROA SDROE M_LLP Z ZP ZP1 ZP2 M_ROA M_ROE

Sample of large banks (total assets larger than 1 billion €, 152 obs.) CONSTANT

1.2504*** (2.669)

17.294** (2.390)

0.9941 (1.234)

5.5052*** (3.339)

65.888** (2.198)

7.0737* (1.820)

58.814** (2.245)

0.8915*** (3.594)

3.8030 (0.776)

FAMILY

-0.0037* (-1.879)

-0.0800** (-2.339)

-0.0037 (-1.336)

0.0941 (0.973)

-0.0361 (-0.243)

-0.0268 (-1.591)

-0.0092 (-0.069)

0.0006 (0.330)

0.0075 (0.225)

INSTITUT

0.0020 (0.998)

0.0737* (1.786)

0.0037 (1.285)

-0.4256** (-2.959)

-0.4798*** (-3.068)

-0.0603** (-2.452)

-0.4194*** (-3.153)

0.0007 (0.361)

0.0368 (0.908)

COMPANY

0.0028 (1.146)

0.1200 (1.106)

0.0045 (0.767)

-0.2205* (-1.889)

-0.2912** (-2.205)

-0.0354* (-1.895)

-0.2557** (-2.225)

-0.0038 (-1.063)

-0.1237 (-1.143)

M_LNTA

-0.0476** (-2.011)

-0.5703 (-1.408)

-0.0444 (-0.769)

-1.6748* (-1.679)

-1.3597 (-0.805)

-0.0862 (-0.347)

-1.2734 (-0.873)

-0.0053 (-0.314)

0.5369 (1.587)

M_OEQUITY

0.0365*** (4.966)

-0.5703 (-1.408)

-0.0572** (-1.115)

0.7815 (1.028)

0.6066 (0.695)

0.0365 (0.287)

0.5700 (0.758)

0.0646*** (7.242)

0.0781 (0.439)

M_DEPOSIT

-0.0022 (-0.565)

-0.2026 (-1.292)

-0.0091** (-2.565)

0.1097 (0.623)

-0.2004 (-0.847)

-0.0544* (-1.835)

-0.1460 (-0.692)

-0.0027 (-0.937)

-0.0157 (-0.382)

LISTED

-0.2563** (-2.588)

-5.2011* (-2.215)

-0.0087 (-0.988)

12.4628 (1.612)

17.859 (1.360)

4.1966** (1.984)

13.662 (1.221)

0.2695** (2.406)

6.1058*** (2.644)

R² 0.1885 0.0962 0.0500 0.0858 0.0501 0.0644 0.0488 0.2780 0.1441 Sample of small banks (total assets smaller than 1 billion €,97 obs.)

CONSTANT 0.8546*** (3.102)

5.4652*** (3.080)

-0.1044 (-0.083)

7.5472*** (3.646)

77.424** (2.000)

7.8741* (1.969)

69.550* (1.905)

0.2411 (0.317)

3.6301 (0.988)

FAMILY

-0.0072 (-1.254)

-0.0810* (-1.743)

-0.0104 (-0.755)

1.0274 (0.921)

0.5168 (0.508)

0.0419 (0.567)

0.4749 (0.500)

-0.0001 (-0.037)

0.0129 (0.350)

INSTITUT

0.0045 (1.385)

0.0442* (1.818)

0.0242* (1.831)

-0.2190 (-1.500)

-0.4173** (-2.281)

-0.0310 (-1.439)

-0.3863** (-2.300)

0.0073 (1.122)

-0.0061 (-0.174)

COMPANY

-0.0057* (-1.950)

0.0079 (0.907)

0.0079 (0.907)

0.1200 (0.663)

-0.1240 (-0.572)

-0.0232 (-0.942)

-0.1007 (-0.505)

0.0009 (0.169)

0.0105 (0.455)

M_LNTA

-0.0248 (-1.126)

-0.0253 (-0.166)

0.0981 (0.640)

-2.9527 (-1.296)

-0.6751 (-0.159)

-0.1800 (-0.897)

-0.4950 (-0.118)

0.0416 (0.741)

0.5510** (2.037)

M_OEQUITY

0.0636*** (6.531)

-0.0071 (-0.228)

-0.0760* (-1.872)

0.1575 (0.425)

-0.6459 (-0.992)

-0.0680 (-1.608)

-0.5778 (-0.917)

0.1148** (2.589)

-0.0220 (0.242)

M_DEPOSIT

-0.0029 (-1.097)

-0.0105 (-0.524)

-0.0360 (-1.147)

0.1083 (0.336)

0.1351 (0.309)

-0.0663* (-1.681)

0.2015 (0.482)

-0.0083 (-1.268)

-0.0604* (-1.760)

LISTED

-0.3321* (-1.866)

1.0145 (0.778)

3.5426 (1.439)

-0.4.526* (1.946)

-47.923** (-2.211)

1.0184 (1.291)

-48.941** (-2.274)

1.1104** (2.500)

9.5476*** (4.137)

R² 0.5230 0.1184 0.0721 0.070 0.0526 0.0898 0.0526 0.5123 0.1460

.***, ** and * indicate significance, respectively, at the 1%, 5% and 10%levels respectively. t-statistics are corrected for heteroskedasticity following White’s methodology. Variable definitions (standard deviations and means are computed over the period 1999-2005): SDROA= standard deviation of the return on average assets; SDROE = standard deviation of the return on average equity , M_LLP = mean of the ratio loan loss provision to net loan; Z = Z-score; ZP = ZP-score; ZP1=measure of bank portfolio risk; ZP2 = measure of leverage risk; M_ROA= mean of the return on average asset; M_ROE= mean of the return on average equity; M_LNTA= mean of logarithm of total asset; M_OEQUITY = mean of the ratio equity to total assets orthogonalized with TA; M_DEPOSIT = mean of the ratio of deposits to total assets. The variables FAMILY, INSTITUT and COMPANY represent the percentage of stock held respectively by families and individuals, institutional investors and non-financial companies.

27

Table A4. Influence of the nature of the main shareholder on risk, probability of default, and profitability (Model 1), cross-section OLS regressions Risk measures Default Risk measures Profitability measures SDROA SDROE M_LLP Z ZP ZP1 ZP2 M_ROA M_ROE

CONSTANT

0.0921*** (3.618)

9.2539*** (2.771)

0.7588 (1.020)

6.3582 (5.028)

76.813*** (3.982)

6.5531**** (2.750)

70.260*** (4.037)

0.6103* (1.885)

3.8282 (1.468)

FAMILY_OWNED

-0.1688 (-0.889)

-2.6473 (-0.869)

-0.5588** (-2.106)

3.4500 (1.260)

-14.064 (-1.100)

-2.4133** (-2.163)

-11.651 (-0.968)

0.1659 (0.708)

1.6046 (0.522)

INSTITUT_OWNED

0.2506 (1.276)

2.5488 (1.303)

1.3411* (1.915)

-14.434* (-1.681)

-34.298*** (-4.147)

-3.3123*** (-3.216)

-30.985*** (-4.161)

0.3431 (0.829)

0.0981 (0.044)

COMPANY_OWNED

-0.0576 (-0.327)

4.8497 (0.869)

0.1810 (0.472)

5.2993 (0.522)

-16.367 (-1.595)

-2.5221** (-2.420)

-13.845 (-1.443)

-0.3646 (-1.176)

-6.3551 (-1.191)

M_LNTA

-0.0200 (-1.540)

-0.0978 (-0.557)

-0.0003 (-0.005)

-2.2966** (-2.200)

-1.8333 (-1.313)

-0.0484 (-0.335)

-1.7849 (-1.362)

0.0219 (1.240)

0.5465***

(3.188)

M_OEQUITY

0.0541*** (7.296)

-0.1004 (-1.460)

-0.0492*** (-2.617)

0.0207 (0.071)

-0.2869 (-0.886)

-0.0576 (-1.323)

-0.2293 (-0.791)

0.1009***

(3.210)

0.0311 (0.402)

M_DEPOSIT

-0.0026 (-1.043)

-0.0302 (-0.924)

-0.0235 (-1.512)

0.0863 (0.524)

-0.0828 (-0.316)

-0.0643** (-2.745)

0.0895 (-0.086)

-0.0067* (-1.836)

-0.0497* (-1.817)

LISTED

-0.1615** (-2.076)

-1.8055 (-1.524)

1.1661 (1.411)

-3.0998 (-0.344)

-6.5905 (-0.496)

3.6671** (2.418)

-10.257 (-0.831)

0.6172***

(3.713)

7.3780***

(6.109) Number of obs. 241 241 239 249 241 241 241 241 241 R² 0.2934 0.0326 0.0440 0.0333 0.0213 0.0639 0.0216 0.4327 0.1006 .***, ** and * indicate significance, respectively, at the 1%, 5% and 10%levels respectively. t-statistics are corrected for heteroskedasticity following White’s methodology. Variable definitions (standard deviations and means are computed over the period 1999-2005): SDROA= standard deviation of the return on average assets; SDROE = standard deviation of the return on average equity , M_LLP = mean of the ratio loan loss provision to net loan; Z = Z-score; ZP = ZP-score; ZP1=measure of bank portfolio risk; ZP2 = measure of leverage risk; M_ROA= mean of the return on average asset; M_ROE= mean of the return on average equity; M_LNTA= mean of logarithm of total asset; M_OEQUITY = mean of the ratio equity to total assets orthogonalized with TA; M_DEPOSIT = mean of the ratio of deposits to total assets; LISTED= dummy variable that equals 1 if the bank is listed in a stock market, and 0 otherwise. Family-owned, institutional investor-owned, and company-owned are dummy variables which takes the value of one when ownership is at least equal to 33% of the equity and 0 otherwise.

28

Appendix 2 Our Model 1 is defined as:

5

i 0 j ji 6 i ij=1

RISK C Z= α + α +α + ε∑

with C1i = MANAGERi ; C2i= FAMILYi ; C3i= INSTITUTi ; C4i= COMPANYi; C5i= BANKi

and Zi is a vector of control variables.

As we have 4

5i jij=1

C =100- C∑ , we can rewrite Model 1 as following17:

4 4

i 0 j ji 5 ji 6 i ij=1 j=1

4

0 5 j 5 ji 6 i ij=1

RISK C + (100 C ) + Z

( 100 ) ( )C Z

= α + α α − α + ε

= α + α + α −α +α + ε

∑ ∑

∑

We can then estimate the following Model:

4

' 'i 0 j ji 6 i i

j=1RISK C Z= α + α +α + ε∑

with ' '0 0 5 j j 5α = α +100α and α =α -α , j=1,..,4 .

The estimated coefficient associated to each ownership component Cj has to be interpreted as

the effect of a substitution between this component and the component C5.

Regarding Model 2, we have

5 5

i 0 j ji j ji i i ij=1 j=1

RISK C C *LISTED + Z= β + β + γ λ + ε∑ ∑

We can rewrite Model 2 as following:

4 4 4 4

i 0 j ji 5 ji j ji i 5 ji i i ij=1 j=1 j=1 j=1

4 4

0 5 5 i j 5 ji j 5 ji i i ij=1 j=1

RISK C + (100 C ) + C *LISTED + (100 C )*LISTED + Z

( 100 100 *LISTED ) ( - )C ( - )C *LISTED + Z

= β + β β − γ γ − λ + ε

= β + β + γ + β β + γ γ λ + ε

∑ ∑ ∑ ∑

∑ ∑

17 For 198 out of the 249 banks, the sum of the percentages of equity held by our five categories of shareholders is equal to 100%.