1
Master Thesis
How does privatization affect the risk-taking behavior
of privatized banks?
WanYun Chen
July 12, 2013
University of Groningen: S2017482 Uppsala University:820430-P173
Msc International Financial Management Faculty of Economics and Business
Msc Business and Economics Faculty of Social Science
2
Abstract
This paper examines the risk-taking behavior of privatized banks pre- and post- privatization. Although previous scholars suggest that the post privatization bank performance improve, the studies on the risk-taking behavior of privatized banks are merely developed and ambiguous. The aim of this paper is to investigate the short-term and long-short-term relationship between privatization and the change of risk-taking behavior of privatized banks, and to examine how the country and the firm level characteristics or the attributes of privatization affect the risk-taking behavior of privatized banks. I expect that the privatized banks experience a significant decrease in risk-taking behavior after privatization. Further, the foreign ownership and the share issue privatization also matter in explaining the risk-taking behavior of privatized banks. This study intends to shed a light that successful privatization program can benefit the privatized banks with lower risk, and the country and the firm level characteristics will also affect the risk following privatization.
JEL classification: G21, G32
3
1. Introduction
Large segments of the global banking system have been transferred from state to private hands over the past two decades, and much more is planned to be sold in the near future (Megginson, 2005). What has caused the relinquishing of government control for the banking sector? The two important factors have stood out. First, the empirical evidences have shown that the state ownership was not working well as planned. La Porta et al, (2001) document that the government-owned banking institutions usually had the poor performance than private owned banking firms did, especially in developing countries. Secondly, the financial system plays an important role in society. Further, the studies show that an efficient and stable financial system promotes the economic growth (Megginson, 2005). Thus, the government believes that the government-owned banking institutions’ efficiency can be improved through privatization and then the more efficient and stable financial system helps the economic growth.
In the past, the state-owned banks used to be the government’s main tool to support the policies and fund the specific industries. Thus, the state-owned banks cannot operate as profit-oriented as private owned banks can. Moreover, some inappropriate government policies may lead to the poor loan quality and the higher default risk. So, the government believes that relinquishing the control of state-owned banks benefits bank performance and contributes to a developed financial system which finally will stimulate the economic growth (Andrews, 2005).
4 always the favorite tools for governments to fund the specific sectors on favorable terms, and it makes state-owned banks more opaque, especially compared to the private commercial banks in public stock market.
One is the limitation of abilities to analyze the privatized bank data. Based on the backgrounds, researchers usually focus on the countries that they are familiar with. It means that they have more knowledge about the backgrounds of privatized banks, the history of financial liberation, the government policies of privatization, and even the language. Thus, they can provide the comprehensive analyses of these privatized banks within a single country.
However, in multiple countries content, it is difficult to handle so many differences from the country and the firm level characteristics because these characteristics may interact with each other and influence the interpretations of the privatization results. Although it is more difficult, it is still necessary to conduct a privatization research in multiple countries because when taking the country and the firm level differences into account, the better understanding of privatization cab be provided by this kind of research.
According to Megginson’s (2005) literature review, intensive studies document that the financial performance (especially profit) enhances significantly for non-financial firms, but relatively few empirical analyses of bank privatizations have been generated, and most of these have appeared only very recently.
Though empirical analyses of bank privatizations have been generated, most of them only devoted attention to the argument of financial and operating performance improvement after privatization in the banking industry (N. Boubakri et al., 2005). For example, Verbrugge et al. (1999) suggest that limited bank performance improvements exhibit after privatization, but Bonin et al. (2005) argue that financial performance is significantly improved. Moreover, a study by the International Monetary Fund (2000) mention that profitability (return on equity) of foreign banks is significantly higher than that of domestic banks in transition economies.
5 developed and developing countries. He captures the decrease of risk-taking behavior following privatization.
Even though these studies make no explicit predictions about the changes of risk-taking behavior, traditional theoretical banking scholars suggest that the increase in competition encourages banks to take excessive risk (Dermine, 1986). Furthermore, empirical evidence on bank privatization supports the competition effects hypothesis which states that privatization increases competition and causes the negative stock reactions of rival banks.
On the other hand, the empirical study demonstrates that the risk-taking behavior of privatized banks decreases because the new foreign owners execute the banking business more efficient and risk-controlled (Boubakri et al., 2005).Some empirical evidences also support that to relinquish government control of state-owned banks will improve the loan quality and reduce the credit risk (Lin and Zhang, 2009). A considerable amount of literature has been published on bank risk, but there is less evidence on investigating the relationship between privatization and bank risk. Most of the studies that related to bank risk focus on the moral hazard problem. So, I summarize the prior researches on privatization and risk as follows. Boubakri et al. (2005) find that the ownership type and industry concentration significantly affect the risk-taking behavior. Banks controlled by industrial groups take the highest risk exposure, while foreign-owned banks take the least exposure. Further, Mohsni and Otchere (2011) did a great work in examining the relationship between risk-taking and privatization, but his work more focused on the comparison of the change of risk-taking behavior between privatized banks and rival banks.
To expand the understanding this topic, my research follows his spirit. Instead of focusing on the comparison of the change of risk-taking between privatized banks and rival banks, I concentrate on the privatized banks themselves and include more the country and the firm level characteristics in this research. The short-term and long-term research periods are both employed in my study. In that case, the short-long-term and long-term effects of privatization on risk-taking behavior would be represented in this research.
6 number of privatization transactions, the foreign ownership, and the share issue privatization). Third, the different research time periods and more variables of the country and the firm level characteristics are employed in this paper, when examining the change of risk-taking behavior following privatization.
Hereby, the research questions are as follows.
Does the risk-taking behavior of privatized banks decrease after privatization? How does the foreign ownership affect the risk-taking behavior of privatized banks? How does the share issue privatization (SIP) affect the risk-taking behavior of privatized banks?
How does the number of privatization transactions affect the risk-taking behavior of privatized banks?
In this paper, I pursue three goals: First, I use an unbalanced panel data of bank information from different developing and developed countries to examine the impact of privatization on risk-taking behavior. Secondly, I further assess the relationship between risk-taking behavior of privatized banks and the country and the firm level characteristics following privatization. Thirdly, using the short-term and long-term research periods, I observe the short-term and long-term effect of privatization on the risk-taking behavior of privatized banks.
In sum, my paper contributes to the little literature that focuses on and thoroughly examines the risk-taking behavior of privatized banks following privatization. I extend my sample to privatized banks in developed and developing countries and focus on the sample itself, not the comparison with rival banks. The study sheds a light on how the country and the firm level characteristics affect the risk-taking behavior of privatized banks following privatization. Additionally, it provides the short-term effect (three years) and long-term trend (1998-2011) of privatization on risk. I attempt to extend the research scope by different time frames and different variables in this paper.
7
2. Literature Review
In section 2, I will summarize the literature concerning privatization and risk. First, I will address the importance of privatization and conclude the empirical evidence on privatization and risk. Secondly, I will introduce the competition effects hypothesis from the traditional banking literature. Finally, I will develop the hypotheses based on the theory and my argument.
2.1 The importance of privatization and its impact
Increasing evidence on the costs of public ownership highlights the need for bank privatization and the potential benefits of shifting to private ownership, especially in low income countries where state ownership is high (La Porta et al., 2002). While privatization of banks in developed countries was initiated as early as the mid-1980s, most developing countries started selling their banks a decade later.
Privatization seems like a panacea for all the state-owned enterprises’ inefficiency and underperformance when the government faces the public pressure to improve the performance of the state-owned enterprises. However, to relinquish the control, or not to relinquish the control is a question. Most of the state-owned enterprises have their own specific purposes when they are set up by government in the beginning. The purposes are to help the social welfare, to keep the stability of society, and to support some specific industries. For the government, to relinquish the control of the state-owned enterprises indicates that the state-state-owned enterprises will not fulfill the purposes well after being privatized.
On the other hand, the government will also face the pressure to improve the performance of the state-owned enterprises if the government does not relinquish the control of the state-owned enterprises as the public expect.
Privatization indeed brings some benefits, but also may cause some serious problems. For example, selling the shares or assets of the state-owned enterprises may improve the efficiency and performance or provide the government supplemental funds.
8 shares or assets of state-owned enterprises more easily. The privatization itself may also change the roles of state-owned enterprises, and then bring the negative influence to the society. Take the state-owned banks for instance. The state-owned banks were set up to operate for financial stability and support some specific industries. When banks have been privatized, they turn to be profit-oriented and stop to lend some specific industries that they supposed to do. This change will encourage the privatized banks to pursue more profits by taking more risk. If the risk-taking behavior of privatized banks is not regulated and monitored appropriately, the increasing risk will cause the instability of whole financial system and lead to a financial crisis.
That is why the study of privatization is so important. This kind of research not only sheds a light in the existing academic literature, but also supports the policy makers to carry out the privatization with appropriate regulatory system.
2.2 The empirical evidence on privatization and risk
9 Otchere (2005), conducts a comparable pre- versus post-privatization study of operating performance changes, and also examines stock price performance, for 21 bank privatizations in low- and middle-income countries from 1989 to 1997. He employs the same measures for 28 rival firms for comparison purposes, and represents that privatization announcements cause significantly negative stock price reactions from rival banks. These findings support the ‘‘competitive effects’’ hypothesis; and Omran (2004) who examines the financial performance of privatized Egyptian banks and changes in capital and liquidity risk but generally observes insignificant results.
More recently, Mohsni and Otchere (2011) demonstrate that prior to privatization, privatized banks exhibit higher risks than rivals; following privatization, privatized banks undergo a significant decrease in accounting risk measures and exhibit lower idiosyncratic and total market risk.
In sum, the empirical evidence of privatization shows that the risk-taking behavior of privatized banks decreases and the competition increases following privatization. These findings also indicate that the competition effects hypothesis exists. More empirical evidence concerning the presence of the competition effects hypothesis will be addressed in the section 2.3.
2.3 The competition effects hypothesis and the traditional banking literature
10 There is no doubt that the empirical evidence shows the presence of competition effects following privatization, but the relationship between competition and risk is reported differently in empirical research and traditional banking theory.
A significant reduction of risk-taking behavior of privatized banks and the presence of competition effects are both supported by the empirical studies, but the traditional banking scholars supports the competition effects hypothesis and concludes that banks usually choose more risky portfolios when confronted with increased competition (Boyd et al., 2004; Marcus, 1984; Dermine, 1986; Chan, Greenbaum and Thakor ,1986).
The traditional banking literature views bank risk-taking behavior within a moral hazard framework where deposit insurance and the value of the bank charter are the main drivers. This literature represents that deposit insurance provides the bank an incentive to intentionally take on risk of failure, possibly without limit. The deposit insurance creates a payoff structure that large gains go to bank shareholders and large losses to the government. Thus, within this model, when the privatization increases the competition of banking sector, the privatized banks would take more risk than before. Moreover, Keeley (1990) find that the positive relationship between competition and risk-taking behavior shows because an increase in competition causes bank charter value to decline, erodes monopoly rents and enhances bank risk-taking. While traditional banking scholars claim the presence of a positive relationship between competition and risk-taking behavior, little research has been done to explore how privatization affects the risk-taking behavior of banks in a framework of heightened competition. Based on the traditional banking literature, the impact of privatization on the risk-taking behavior of privatized banks remains empirically and theoretically ambiguous.
2.4 The argument about the change of bank risk-taking behavior following privatization
11 with private banks, they are expected to be less risky and should see their risk increase if they become privatized.
On the other hand, it could also be argued that state-owned banks usually tend to extend credits for political reasons with little economic justification, and lend riskier segments at below market rates as the subsidy, and thus expected to be more risky than similar private entities and should see their risk decrease if they become privatized.
As the related theories and argument above, the change of risk-taking behavior following privatization is still lack of explicit prediction. The privatized banks become more risky when the privatized banks face the heightened competition caused by privatization and have more pressure to pursue profits. On the contrary, the privatized banks are less controlled by government so the prudent lending leads to the decrease of risk-taking. Based on the previous empirical evidence, I argue that the privatized banks will experience the reduction of risk when they are not served as government’s tool, even though the competition effect exists. All in all, I develop my hypothesis 1 as follows.
H1: The risk-taking behavior of privatized banks in post privatization is lower than that in pre privatization.
2.5 Risk-taking behavior and foreign ownership
A firm ownership structure can be defined along two dimensions. First, the degree of ownership concentration: firms may differ because their ownership in more or less dispersed. Secondly, the nature of the owners: given the same degree of concentration, two firms may differ if the government holds a majority stake in one of them; similarly, a stock firm with dispersed ownership is different from a mutual firm. According to Levine (1997), the ownership structure of banks is a crucial variable in the process of financial deepening and economic growth. Banks are not only providing the much needed financing for restructuring, but also are expected to play a major monitoring role, especially in institutionally weak environments.
12 performance in transition economies. They find that privatization to foreign owner brings cost advantage in the source of funds for lending.
However, Boubakri et al. (2005) find that post- privatization risk exposure may worsen or improve depending on the type of owner. They find that the risk-taking behavior of privatized banks is affected by ownership type and industry concentration significantly. Privatized banks controlled by industrial groups take the highest risk exposure, while foreign-owned banks take the least exposure.
Similar to Boubakri et al. (2005), Jia (2009) finds that privatization in China has provided incentives for banks to engage in prudent lending. In contrast, Susanto and Rokhim (2011) examine the impact of foreign ownership in Indonesian banking sector after deregulation, and they report that the increasing foreign ownership positively correlates to competition and risk.
It can be concluded that the choice of ownership structure is important in the context of privatization since the set of incentives to which a local investor or a foreign investor may be subjected to, is different, hence potentially yielding divergent privatization outcomes.
Regarding the influence of foreign ownership on the risk-taking behavior of privatized banks, the previous empirical evidence is mixed. So, I select the foreign ownership as the ownership structure variable in the regression model due to its importance in the literature.
I argue that when the privatized banks are partly controlled by foreign ownership, the risk-taking behavior is decreasing because foreign ownership would be more motivated to pursue appropriate risk control and prudent lending. Thus, I develop my hypothesis 2 as follows.
H2: The foreign ownership will decrease the risk-taking behavior of privatized banks.
2.6 Risk-taking behavior and share issue privatization
Based on the literature concerning privatization, the ownership structure and the governance play important roles in affecting post privatization performance. In addition, the method of privatization also affects the privatization outcomes.
13 have higher profit-maximum pressure. This rationale may make privatized banks through SIP lead to higher risk than those through asset sales.
For instance, banks privatized through new share issue are usually subject to higher pressure from financial markets to achieve profit-maximizing results. Such goal can be reached through an increase in inefficiency, and increase in investment in risky assets or a combination of both. Following the rationale above, I develop my hypothesis 3 as follows.
H3: The share issue privatization (SIP) will increase the risk-taking behavior of privatized banks.
2.7 Risk-taking behavior and the number of privatization transactions
When I collect the data of privatization transactions, I find that some state-owned banks are privatized gradually, whereas some are privatized in one time.
Otchere (2009) state that firms that become fully privatized fast are expected to experience a more dramatic shift in their risk behavior than firms that undergo sequemtial privatizations over several years. Cosset et al. (2011) report that the presence of government control is expected to constrain the natural change in risk taking that might ensue from full privatization.
On one hand, it could be argued that state-owned banks are privatized gradually have more time for cultural blending from state owner and private owner, and it will eliminate the difficulties of merge, and thus the risk-taking behavior will decrease gradually.
On the other hand, it could also be argued that state-owned banks are privatized gradually will hinder the managers’ abilities to restructure the banks’ business due to the continued partially government ownership. So, the risk-taking behavior of privatized banks will not decrease as expected.
According to the arguments above, I think that banks are privatized gradually will experience a slight shift in their risk-taking behavior, even though the government control presents. So, I develop the hypothesis as follows.
14
3. Methodology
3.1 Data
3.1.1 Sample and observations
First, I obtain the lists of privatized banks from World Bank Privatization Database and Privatization Barometer Database. The lists include the bank privatization transactions (above 1 million) in developing and developed countries. The transactions of privatized banks developing countries are drawn from World Bank Privatization Database provides, while the transactions of privatized banks developed countries are offered by Privatization Barometer Database.
These banks have been privatized from 2000 to 2008. The total amounts of privatization transactions are 102. Some state-owned banks have been privatized by multiple privatization transactions, so the sample banks are less than 102. To identify sample banks and collect the necessary information, I retrieve all the banks on Bankscope. Then, I obtain an unbalanced sample of 86 banks from 37 countries with annual data for the period 1998-2011. The list of privatized banks is presented in Appendix1. The sample yields a total of 1204 observations. Since not all variables are available for all the banks, fewer observations are included in some of regressions. I also use World Bank Database to collect annual information on GDP per capita and GDP growth rate from 1998 to 2011. The website of Central Intelligence Agency is employed to provide the common law countries list. 1
The vectors of bank risk are the Z-score, the volatility of ROA, the ratio of impaired loans to gross loans, and the volatility of ROE. For each bank, I use accounting measures of risk around the year of privatization which I define as follows: it is the year on which the government divested some (or all) of its stake in the bank for the first time. The research period of risk measurement is 3 years before (year: -3 to -1) and after privatization (year: +1 to +3). It depends on which year that the privatization occurred (year: 0).The privatization bank transactions are drawn from 2000 to 2008, and the data of bank years would be collected from 1998 to 2011 due to data availability of Bankscope, instead of 1997 to 2011. Besides the short-term
1
15 research period, I also employ the long-term research period as 1998 to 2011 to run the regression models with year fixed effects.
I conclude the distributions of the observations in as Table 1. This table presents the summary for the sample of privatizations in pre and post three years. Panel A and Panel B demonstrate the frequency distribution of the sample of privatizations. Panel C presents the distributions of the privatizations by method, and Panel D shows the distributions of foreign ownership for privatization transactions.
As Panel A shows, the total amounts of privatization transactions are 102. According to the amounts of privatization transactions implemented, the first four countries are China, Italy, Russian Federation and Pakistan. As Panel B presents, yearly frequency distributions are almost the same in every year, besides 2008.
Table 1
Description of the sample
Panel A: Distributions of privatization transactions by country
Country No. of privatization transactions Country No. of privatization transactions
AL: Albania 1 LV: Latvia 1
AT: Austria 1 LY: Libya 1
BD: Bangladesh 2 NG: Nigeria 1
BG: Bulgaria 2 PH : Philippines 1
CN: China 16 PK: Pakistan 8
CZ: Czech Republic 1 PL: Poland 2
DE: Germany 4 PT: Portugal 2
EG: Egypt 1 RO: Romania 2
ES: Spain 1 RS: Serbia 1
FI: Finland 4 RU: Russian Federation 6
FR: France 2 SE: Sweden 1
GR: Greece 6 SI: Slovenia 1
HR: Croatia 2 SK: Slovak Republic 2
HU: Hungary 4 TH: Thailand 1
ID : Indonesia 6 TN: Tunisia 1
IE : Ireland 2 TR: Turkey 1
IT: Italy 11 TZ: Tanzania 1
KZ: Kazakhstan 1 UY: Uruguay 1
LB: Lebanon 1 Total 102
16 Table 1
Description of the sample (continued)
Panel B: Distributions of privatization transactions by year
Year Yearly frequency distributions
2000 15 2001 18 2002 10 2003 13 2004 10 2005 12 2006 11 2007 12 2008 1 Total 102
Panel C: Distributions of privatization transactions by method Share issue privatization(SIP) 38
Private asset sale 64
Total 102
Panel D: Distributions of foreign ownership for privatization transactions Foreign ownership (>30%) 57
Non Foreign ownership 45
Total 102
Panel C represents that privatization transactions through private asset sales are more than though SIP in this sample. Panel D reveals the distributions of foreign ownership for privatization transactions. In 102 privatization transactions, only the privatized banks belong to the 57 privatization transactions have more than 30% foreign ownership.
3.2 Methodology
17 I conduct the regression models in the same spirit of Boubakri et al (2005), and Mohsni and Otchere (2011). The analyses focus in the effects of privatization on bank risk, and also examine the relationship between privatized bank risk and the country and the firm level characteristics. The estimation method is least squares, and the year fixed effects specification is employed in all the regressions. In addition, the foreign ownership dummy variable following the methodology proposed by Lin and Zhang (2009), I evaluate the selection effects associated with foreign ownership changes.
3.3 Models and variables 3.3.1 Models
In this part, I introduce all the models that will be employed and the steps of analysis. My research consists of the data with cross-section and time series characteristics, panel regression will be used. To investigate the effect that privatization and the firm level characteristics have on bank risk behavior, I use multivariate panel regression which allows multiple variables into the model for explanation. In order to estimate the relationship more precisely, the panel regression with time-period fixed effects will be used, and I assume each year has a unique intercept.
In the multivariate analysis, the paper estimates the following general models. For each risk proxy, I will develop 3 regression models through the steps. Following the methodology proposed by Mohsni and Otchere (2011), I will run the regressions of the Z-score and the volatility of ROA as main empirical results, and run the regressions of the ratio of impaired loans to gross loans and the volatility of ROE as robustness check. All the regressions will apply short- term research period (3 years pre- and post- privatization) and long-term research period (1998-2011) to capture the short-term effect and long-term trend.
Step 1: For privatization
Step 2: Adding country level variables
18 Step 3: Adding the firm level variables
Where is one of the four risk measurements of each bank i at time t, namely:
the z-score, the volatility of ROA, the ratio of impaired loans to gross loans and the volatility of ROE. The variables are divided into three categories: the variables of privatization attributes, the variables of country level characteristics and the variables of firm level characteristics. To better understand, the variables are summarized in Table 2. Table 2 contains the description of all the variables used in regression models. I will treat all the variables thoroughly in the following paragraphs.
Table 2
Variables employed in regression models
Symbol Definition Source
one of the four risk measurements of each bank i at time t
The Z-score The natural logarithm of (the return on assets plus the equity-to-asset ratio, divided by the standard deviation of return on assets)
Author’s calculation based on the data from Bankscope
The volatility of ROA
The volatility of return on assets is calculated as the standard deviation of ROA for three years beforehand
Author’s calculation based on the data from Bankscope
The ratio of impaired loans to gross loans
The ratio of impaired loans to gross loans
Bankscope
The volatility of ROE
The volatility of return on equity is calculated as the standard deviation of ROE for three years beforehand
Author’s calculation based on the data from Bankscope
19 Table 2 (continued)
Symbol Definition Source
Pre_Post A variable equals to years before or after privatization
Author’s calculation based on the data from World Bank Privatization Database and Privatization Barometer Database
PV_times A variable equals to the number of privatization transactions
World Bank Privatization Database and Privatization Barometer Database For a dummy variable, which equals
to 1 if the foreign ownership is higher than 30%
Bankscope
SIP a dummy variable, which equals to 1 if the bank has been
privatized by share issue privatization
World Bank Privatization Database and Privatization Barometer Database
Developing a dummy variable, which equals to 1 if the bank is from a
developing country and 0 otherwise
World Bank Privatization Database and Privatization Barometer Database
Law a dummy variable, which equals to 1 if the bank is from a
common law country and 0 otherwise
The website of Central Intelligence Agency
GDP Growth GDP growth rate World Bank database
GDPPC GDP per capita World Bank database
COI Cost to income ratio Bankscope
Leverage The ratio of equity to liability Bankscope Size The natural logarithm of total
assets
Author’s calculation based on the data from Bankscope
20 3.3.2 Risk measures: dependent variables
Four accounting-based risk measures are used to examine the effect of privatization on the risk-taking behavior of privatized banks in this paper. The four accounting-based measures of bank risk are the z-score, the volatility of ROA, the ratio of impaired loans to gross loans and the volatility of ROE. The description of each measure is as follows.
The z-score: A popular risk measure in the banking and financial stability related literature that reflects a bank’s probability of insolvency is the Z-score. It widespread use is due to its relative simplicity and the fact that it can be calculated by only accounting information (Strobel, 2013). The z-score is defined as the inverse of the probability of insolvency and is estimated as the return on the assets plus the equity-to-asset ratio, divided by the standard deviation of the return on assets. The z-score measures the distance from insolvency (Roy, 1952). I use the natural logarithm of the z-score, which is less skewed and follows the normal distribution for the analysis. A higher z-score (lower risk) indicates that the bank is more stable.
The volatility of ROA: It is calculated as the standard deviation of the return on assets for three years beforehand. This paper uses a three-year moving window to calculate the volatility of ROA.
The ratio of impaired loans to gross loans: It is defined as the ratio of impaired loans to gross loans. This ratio approximates a bank’s exposure to credit risk. Barth et al. (2002) and Gonzalez (2005) use similar ratios to measure bank risk. A higher ratio indicates an increase in the credit risk exposure.
The volatility of ROE: Similar to ROA, it is calculated as the standard deviation of the return on equity for three years beforehand. This paper uses a three-year moving window to calculate the volatility of ROE.
3.3.3 The variables of privatization attributes, country and firm Level characteristics: independent variables
3.3.3.1 The variables of privatization attributes
PRE_POST: A variable is equal to the years before or after privatization. For
21 calculation. The inclusion of this variable allows us to examine the trend of changes in risk-taking behavior of privatized banks before and after privatization.
PV_times: A variable refers to the number of privatization transactions that each
bank had from 2000 to 2008. For example, if a bank has been privatized in 2000 and 2002, the PV_times of year 1998-2001 equals 1, while the PV_times of year 2002-2008 equals 2.
For: A dummy variable is equal to 1 if the bank has higher than 30% foreign
ownership and 0 otherwise. Foreign investors are major stakeholders in the privatization process especially in developing countries. Cosset et al., (2007) find that the presence of foreign investors as owners of the privatized banks leads to the higher efficiency and to an increase in risk-taking.
SIP: A dummy variable equals 1 if the state-owned bank has been privatized
through a share issue privatization (SIP) and 0 if it has been privatized through a private asset sale. The privatized banks through share issue privatization may have higher market pressure to pursue maximum profit and thus are expected to increase the risk following privatization.
3.3.3.2 The variables of country level characteristics
Developing: A dummy variable is equal to 1 if the privatized bank is from a
developing country and 0 otherwise. I examine how the change in risk-taking behavior of privatized banks following privatization differs between developed and developing countries. Banks operating in developing countries are prone to information and accounting problems that may cause excessive risk-taking behavior (Beck et al., 2005).
Law: A dummy variable equals 1 if the privatized bank belongs to a common law
country and 0 otherwise. Common law countries are more likely to provide shareholders with better protection than are those of French civil law (La Porta et al., 1997), therefore it is expected that the privatizations in common law countries would lead to the higher risk-taking behavior than the privatizations in other legal systems. The dummy variable, Law, is then expected to have a positive sign.
GDP Growth: A variable captures the economic growth of the country that the
22 growth rate, so the privatized banks in higher GDP growth countries may cause higher risk. GDP Growth is expected to have a positive sign.
GDPPC: I use the natural logarithm of GDP per capita to capture the level of
development of the economy. The higher GDP per capita implies the higher development of the economy, so the higher GDP per capita country is expected to provide the privatized bank the more developed and transparent financial environment. In this more regulated financial environment, the risk-taking behavior of privatized bank is expected to be well controlled and monitored. Therefore, GDPPC is expected to have a negative sign.
3.3.3.3 The variables of firm level characteristics
COI: A variable is equal to the ratio of cost to income ratio. A positive effect of
inefficiency on risk-taking was found (Kwan and Eisenbeis, 1997). COI captures the efficiency of the privatized bank. Kwan and Eisenbeis (1997) suggest that the lower efficiency level cause greater future risk and the efficiency improvements tend to shore up bank capital positions. They also find that inefficiency leads to a higher probability of default. COI is therefore expected to have a positive sign
Leverage: A variable is equal to the leverage ratio of bank, and it is calculated as
the ration of equity to liability. Following Lev (1974), higher leveraged firms tend to exhibit greater risk. To control this leverage effect, I use the book value of equity to liability ratio in year t-1, and this leverage information is drawn from Bankscope. A higher equity to liability ratio should be negatively correlated to bank risk.
Size: Larger bank sizes are usually more stable and show little information
asymmetry; therefore they have a higher potential to diversify away their risk. This variable is measured as the natural logarithm of total assets and is expected to have a negative relationship with bank risk.
ROE: ROE implies the profitability of the bank. It is calculated as the return on
equity. Higher ROE can be reached through an increase in inefficiency, and increase in investment in risky assets or a combination of both. Rachdi (2011) finds that banks with higher charter value are associated with lower ROE and more bank risk. Thus, ROE is expected to have a negative relationship with risk.
23
4. Empirical Results
4.1 Accounting risk measures in privatized banks 3 years pre- and 3 years post-privatization
First, I start the analysis by examining whether privatized banks experience a change in risk-taking over the three years following their privatizations. To provide a comprehensive assessment of risk-taking behavior, I look at the accounting based proxies of risk. For each proxy, the mean, median, and the standard deviation values for 3 years pre and post privatization are calculated. T-test statistics is used to measure the significance of the differences in mean. In that case, a significant difference in mean can be inferred to the change in risk-taking behavior. The results are presented as Table 3.
Table 3
Accounting risk measures 3 years pre- and 3 years post privatization Privatized banks 3 years post-privatization (year: 1 to year: 3) 3 years pre-privatization (year :-1 to year: -3) Difference (T-statistics) Mean Median Standard
Deviation
24 The table 3 shows the mean, median, standard deviation, and difference in mean test for the volatility of ROE, the volatility of ROA, the ratio of impaired loans to gross loans, and the Z-score.
I find that the three accounting-based measures (the volatility of ROE, the volatility of ROA, and the ratio of impaired loans to gross loans) decrease greatly from pre-privatization period to the post-pre-privatization period at a significance level of 1%. The significant decreases in the three proxies indicate the lower risk-taking of privatized banks. It also indicates an improvement in the bank’s stability and its capital soundness following privatization. The latter may be achieved through an increase in earnings, an increase in capital adequacy ratios, or a mix of both. The other accounting-based risk measure (Z-score) also provide consistent result but not so significant (probability=0.1555) ; that is, the privatized banks experience a reduction in bankruptcy risk following privatization. These results are in line with the hypothesis 1 that the privatized banks have become less risky after privatization.
4.2 Regressions
In this section, I report the results of a multivariate time series analysis of bank risk. The estimations are conducted using least squares method, and year fixed effects is employed in all the regression models.
As section 3.3.1 models stated, for each risk proxy, I will follow the three steps to develop the three specifications. Following Mohsni and Otchere (2011), I will run the regressions of the Z-score and the volatility of ROA to provide the main research results, and run the regressions of the ratio of impaired loans to gross loans and the volatility of ROE as robustness checks.
Take the Z-score as the dependent variable for example; the three specifications are developed as follows. In the first specification, I examine how the risk-taking behavior of privatized banks differs between pre and post privatization with the privatization variable (pre_post), together with the privatization attributes as the number of privatization transactions (PV_times), the foreign ownership (For), and the privatization method (SIP).
25 second specification: a developing dummy variable, a law dummy variable, GDP growth rate and the GDP per capita.
Previous literature on risk-taking behavior of banks highlights the influence of the firm level characteristics such as leverage, size and profitability. Thus, in the third specification, I include these variables and the efficiency variable (COI) to represent the firm level characteristics. The firm level characteristics variables allow me to separate the effect of privatization on the risk-taking behavior of privatized banks from any confounding firm specific effects.
Besides Z-score, each risk proxy (the volatility of ROA, the volatility of ROE, and the ratio of impaired loans to gross loans) as the dependent variable is used in the three specifications respectively. I first examine the effect of privatization on risk-taking behavior in short-term (3 years pre and post privatization), and then use the same privatized banks sample with long research period (from 1998 to 2011) to study the long-term trend of privatization effect. I will discuss about the empirical results of each model in the following sections.
4.2.1 The regressions for 3 years pre and post privatization
I report the results of regressions for 3 years pre and post privatization as Table 4 (next page). The table presents regression results of two accounting-based measures (the Z-score, and the volatility of ROA) on variables that reflect privatization attributes (Pre_post, PV_times, For, SIP), the country level characteristics (Developing, Law, GDPGrowth, GDPPC) and the firm level characteristics (COI, Leverage, Size , ROE). I will analyze the results of the Z-score models and the volatility of ROA models thoroughly.
26 Table 4
Cross-sectional time series regressions results, accounting-based risk measures (3 years pre and post privatization)
Z-score Volatility of ROA
(1) (2) (3) (4) (5) (6) Intercept 0.8281*** (4.8886) 0.2181*** (0.5011) -0.8938 (-1.0168) 1.0478* (1.8642) 3.5665** (2.4819) 3.1027 (1.7980) Pre_Post 0.1127* (1.7341) 0.1358** (2.0756) 0.1052 (1.5947) -0.4336* (-1.9476) -0.5066** (-2.2858) -0.2161* (-1.6151) PV_times -0.2296 (-1.4867) -0.2635* (-1.7087) -0.2056 (-1.4334) 1.0928** 2.1168 1.1944** (2.3347) 0.5450** (1.8659) For 0.3409** (2.2019) 0.5214*** (2.9739) 0.4906*** (2.7916) -0.7128 (-1.3792) -1.4467** (-2.5132) -0.7739 (-2.2465) SIP 0.2362 (1.4914) 0.3522** (2.1397) 0.2281 (1.2690) 0.4829 (0.9015) -0.0571 (-0.1032) -0.1882 (-0.5170) Developing 0.4156 (1.3264) 0.4059 (1.3759) -1.3245 (-1.2871) -0.3841 (1.6578) Law 0.0201 (0.0735) -0.1452 (-0.5376) -0.8746 (-0.9691) -0.1315 (-0.2377) GDP Growth -0.0035 (-1.1149) -0.0044 (-1.4119) 0.0121 (1.1746) 0.0051 (0.8604) GDPPC 0.2489** (2.4873) 0.2564*** (2.7525) -1.0593*** (-3.1614) -0.5527*** (-2.9672) COI -0.0013 (-1.0235) 0.0031 (1.1903) Leverage 0.0051 (1.4984) 0.0172** (2.4449) Size 0.0502 (1.1214) -0.0724 (-0.8186) ROE 0.0288*** (3.9604) -0.0174*** (-2.7787)
Year fixed effects Yes Yes Yes Yes Yes Yes
Observations 281 281 256 302 302 270
0.2438 0.2699 0.3443 0.2057 0.2431 0.3145
27 Thus, the hypothesis 2 that the foreign ownership will decrease the risk-taking behavior of privatized banks is confirmed. The coefficients of SIP and PV_times are not significant in model (1) and (3), so the hypothesis 3 and the hypothesis 4 are not supported.
When the country level characteristics variables are included in the specification (Model (2) and (3)), I find that the effects of a developing country dummy variable and a law dummy variable are mixed and not significant. This finding contradicts to prior researches. The possible rationale is that this research examines the short-term privatization effect (three years), which is shorter than the prior research settings (5 years). So, in the short-term period, the country level characteristics as a developing dummy variable and a common law variable may have the effects of on risk but the effects may not show so strong in statistics.
Moreover, when country level characteristics are included in the specification (Model (2)), the privatization attributes as the number of privatization transactions, SIP and the foreign ownership display more significant influence on risk-taking than the other specifications (Model (1) and (3)).
The coefficient of GDPPC is positive and significant. This finding indicates that privatized banks belongs to a country with higher GDP per capita would have higher z-score (lower risk). This result is also in line with the expectation of the paper. The coefficient of the developing variable is positive as the coefficient of GDPPC, but not significant in this study. This relationship is not consistent with Otchere (2009) ‘s study. Otchere (2009) reports that privatization has encouraged excessive risk-taking among privatized banks in developing countries.
In model (3), profitability has the positive relationship with Z-score (lower risk). This finding is in line with my expectation. I expect the higher ROE accompanies with the lower risk.
28 the number of privatization transactions will decrease the risk-taking of privatized banks. It is reasonable that, in the short run, the volatility of ROA will increase if the bank has been privatized through more transactions. The foreign ownership and SIP do not have significant influences on the volatility of ROA, so the hypothesis 2 and 3 are not supported.
Regarding the country level and the firm level characteristics, the results of model (4)-(6) are similar to those of Z-score models. GDPPC and ROE have negative influences on risk. Moreover, the coefficient of leverage is positive and significant, and it is consistent with the expectation. It means that the higher leverage encourages the risk-taking. Otherwise, the efficiency and size do not affect the risk of privatized banks significantly.
In sum, the results of the volatility of ROA models are mostly consistent with previous studies. The privatization variables (pre_post and pv_times) and firm level characteristics (leverage and ROE) affect the risk-taking behavior of privatized banks most significantly.
4.2.2 The robustness checks of the regressions for 3 years pre and post privatization As a robustness check, I run similar regressions using the ratio of impaired loans to gross loans and the volatility of ROE as risk measures. Regression results are shown in Table 5 (next page). The table presents regression results of accounting-based measures (the ratio of impaired loans to gross loans and the volatility of ROE) on variables that reflect privatization attributes (Pre_post, PV_times, For, SIP), country level characteristics (Developing, Law, GDPGrowth, GDPPC) and firm level characteristics (COI, Leverage, Size , ROE).
I will explain the results of robustness checks thoroughly as follows.
29 Table 5
The robustness checks of the regressions for 3 years pre and post privatization
Impaired loan to Gross Loan Volatility of ROE
(7) (8) (9) (10) (11) (12) Intercept 10.1482*** (9.4650) 11.0383*** (4.3436) 5.7317 (0.9836) 20.2264*** (2.9708) 49.3232*** (2.7911) 28.1843 (0.8552) Pre_Post 0.0308 (0.0735) -0.2463 (-0.5863) -0.1857 (-0.4099) -3.4606 (-1.2836) -3.5221 (-1.2927) -2.5323 (-0.9907) PV_times -0.8863 (-0.8774) -0.4550 (-0.4586) 0.1217 (0.1297) -1.1061 (-0.1768) -0.9462 (-0.1502) -0.0110 (-0.0019) For -2.0305** (-2.1136) -3.2537*** (-3.1543) -5.134*** (-4.6908) 3.9884 (0.6372) -1.5237 (-0.2152) 2.9907 (0.4545) SIP -0.5194 (-0.5079) -1.5749 (-1.4705) -2.2198* (-1.8515) 0.7923 (0.1221) -3.7150 (-0.5459) 0.2359 (0.0339) Developing 1.8310 (0.9387) 3.0629** (1.5213) -9.2336 (-0.7290) 1.6368 (0.1467) Law 6.9486*** (3.4857) 7.1252*** (3.4399) -6.1862 (-0.5564) 11.5115 (1.0892) GDP Growth -0.0259 (-1.4030) -0.0216 (-1.1276) -0.0563 (-0.4419) -0.0393 (-0.3466) GDPPC -0.1700 (-0.2721) -0.7016 (-1.1593) -8.5225** (-2.0671) -3.9747 (-1.1171) COI 0.0170* (1.6709) 0.0395 (0.7761) Leverage 0.0701*** (2.6210) -0.1459 (-1.0838) Size 0.2955 (1.0381) 0.1384 (0.0819) ROE -0.092*** (-4.7160) -0.6020*** (-5.0329) Year fixed effects
Yes Yes Yes Yes Yes
Yes
Observations 256 256 211 301 301 270
0.2829 0.3342 0.4415 0.1400 0.1551 0.2327
30 In Model (8)-(9), I find that the effect of a law dummy variable is positive and significant. This outcome is not found in other risk measure models. It demonstrates that the common law environment has more strong influence on credit risk exposure. As the robustness check, the similar results of leverage and ROE partly verify the results of the Z-score models and the volatility of ROA models.
The volatility of ROE: The volatility of ROE reflects the risk of bank. The higher volatility of ROE indicates the higher risk-taking of bank. In model (10)-(12), most of the coefficients are insignificant, besides the coefficients of GDPPC and ROE variable.
4.2.3 Regressions for long-term trend of privatization effect
In this part, I perform the regressions of the Z-score models and the volatility of ROA models by using the long-term research period data (1998-2011). The table 6 consists of all the regression results of accounting-based measures (the Z-score and the volatility of ROA) on variables that reflect privatization attributes (Pre_post, PV_times, For, SIP) , the country level characteristics (Developing, Law, GDPGrowth, GDPPC) and the firm level characteristics (COI, Leverage, Size, ROE). The table 6 is shown in the next page.
I will address the results of all the regression models thoroughly as follows.
31 Table 6
Cross-sectional time series regressions results, accounting-based risk measures in the long-term period (1998-2011)
Z-score Volatility of ROA
(1) (2) (3) (4) (5) (6) Intercept 0.6359*** (4.5079) -0.5275 (-1.5771) -2.0537*** (-3.2479) 1.4445*** (2.6232) 6.3621*** (4.8513) 4.5448** (2.3811) Pre_Post 0.0435* (1.6865) 0.0565** (2.1958) 0.0671*** (2.6873) 0.1314 (1.2896) 0.1415 (1.4226) 0.0754 (1.0244) PV_times -0.0209 (-0.2187) -0.0118 (-0.1249) -0.0474 (-0.5138) -0.0325 (-0.0833) -0.0026 (-0.0070) -0.0611 (-0.2160) For 0.4882*** (4.3302) 0.3587*** (2.8783) 0.2205* (1.7702) -0.9594** (-2.1611) -1.7821*** (-3.7659) -0.6109* (-1.6663) SIP 0.3829*** (2.9184) 0.4482*** (3.3183) 0.1641 (1.1707) 0.1290 (0.2466) -0.8873* (-1.6750) 0.4457 (1.0555) Developing 0.8577*** (4.1279) 0.7768*** (3.8262) -2.2353*** (-2.8190) -0.9997* (-1.6609) Law 0.1980 (1.0379) 0.0537 (0.2851) -2.5152*** (-3.4907) -1.4601*** (-2.6562) GDP Growth 0.0014 (0.9100) 0.0013 (0.9069) 0.0119* (1.9544) 0.0082* (1.8457) GDPPC 0.2148*** (2.8566) 0.1315* (1.8023) -1.7159*** (-5.8675) -0.8654*** (-3.9348) COI -0.0041*** (-3.1398) 0.0239*** (7.1827) Leverage 0.0109*** (3.2226) -0.0424*** (-4.3704) Size 0.1186*** (3.6684) -0.1545 (-1.5822) ROE 0.0086*** (4.2646) -0.0095*** (-3.0657)
Year fixed effects Yes Yes Yes Yes Yes Yes
Observations 627 627 617 700 700 672
0.1363 0.1664 0.2414 0.0770 0.1383 0.2366
32 The volatility of ROA: Not consistent with the short-term regression results, the coefficient of pre_post variable is not significant in model (4)-(6). However, the foreign ownership has the negative influence on the volatility of ROA significantly. Further, in the short-term period, the risk-taking behavior of privatized banks increases when the banks have been privatized by more transactions, whereas the risk decreases when the banks have been privatized by more transactions in the long-term period. The long-term effect of the number of privatization transactions is in line with my expectation.
I find that the effects of the developing dummy and law dummy are opposite to the previous literature. It shows that the bank from a developing and common law country tends to decrease the risk in the long run. However, the effects of GDP growth and GDP per capita are similar to what I expect. It shows that the bank from a lower GDP growth and higher GDP per capita country tends to decrease the risk-taking. These findings are more consistent with previous literature.
Moreover, when the firm level characteristics are included in the specification (Model (6)), the effects of these variables are more significant than those in the short run. It can be concluded that the risk-taking behavior of privatized banks would be affected by their own firm level characteristics (COI, leverage, and ROE) more. In sum, the results of volatility of ROA models are mostly consistent with previous models. The country and the firm level characteristics affect the risk-taking behavior of privatized banks significantly.
4.2.4 Robustness check of Regressions for long-term trend of privatization effect As a robustness check, I perform the similar regressions by using the ratio of impaired loans to gross loans and the volatility of ROE. Regression results are shown in Table 7. Overall results are consistent with previous findings.
33 Table 7
The robustness checks of the regressions in the long-term period (1998-2011)
Impaired loan to Gross Loan Volatility of ROE
(7) (8) (9) (10) (11) (12) Intercept 8.7970*** (8.6583) 15.3626*** (7.0884) 8.6189** (2.2735) 12.742*** (2.7492) 46.1855*** (4.0983) 13.6639 (0.7438) Pre_Post 0.2087 (1.1464) 0.0693 (0.3817) 0.1562 (1.0377) 1.0968 (1.2800) 1.0393 (1.2154) 0.9424 (1.3299) PV_times -0.4209 (-0.6242) -0.4833 (-0.7369) -0.0397 (-0.0753) -3.0623 (-0.9330) -3.0514 (-0.9354) -2.8943 (-1.0619) For -1.2915* (-1.7123) -2.0527*** (-2.6094) -2.2359*** (-3.3223) -0.4050 (-0.1085) -1.4138 (-0.3478) -1.6571 (-0.4696) SIP -1.5856* (-1.7989) -2.4544*** (-2.7590) -1.5702** (-1.9831) 6.6176 (1.5049) 1.7123 (0.3761) 5.1443 (1.2657) Developing -2.0046 (-1.4771) 0.1946 (1.1701) -22.5624*** (-3.3113) -2.9278 (-0.5054) Law 3.7415*** (2.9166) 4.4492*** (4.2654) -14.4935** (-2.3397) -6.4572 (-1.2206) GDP Growth -0.0204** (-2.0526) -0.0220*** (-2.7198) 0.0625 (1.1873) 0.0237 (0.5516) GDPPC -1.4353*** (-2.8253) -0.5551 (-1.3431) -9.5947*** (-3.8183) -6.6637*** (-3.1515) COI 0.0324*** (4.5588) 0.1686*** (5.2555) Leverage 0.0214 (0.8979) -0.3684*** (-3.9394) Size 0.0017 (0.0095) 0.7149 (0.7603) ROE -0.0303*** (-5.7353) -0.5304*** (-17.6998)
Year fixed effects Yes Yes Yes Yes Yes Yes
Observations 597 597 542 699 699 672
0.2046 0.2550 0.3212 0.1231 0.1446 0.4526
34 The ratio of impaired loans to gross loans: The ratio of impaired loans to gross loans indicates the credit risk of bank. In model (7)-(9), the results are consistent with the short-term regression results. The coefficient of pre_post variable is mixed and not in a significance level. In addition, the coefficients of foreign ownership and SIP variable are negative and significant.
Moreover, in the Model (8)-(9), the coefficients of a law dummy are positive and significant. Thus, this results support that privatized banks in common law would lead to higher risk-taking. It is consistent with the statement that countries of common law legal origin are more likely to provide shareholders with better protection than are those of French civil law origin (La Porta et al., 1997).
In Model (9), the effects of COI and ROE in the long run are the same as those in the short run, but the influence of leverage is not significant as that in previous model.
The volatility of ROE: Dislike the short-term models of the volatility of ROE, most of the coefficients are insignificant, besides the coefficients of GDPPC and ROE variable. The long-term models consist of more variables that have the significant influences.
In model (11), the significant effects of the developing country dummy variable and the law dummy variable are not in line with previous literature, and they are not shown in model (12). The coefficient of GDPPC is followed the expectation. Moreover, when the firm level characteristics are included in model (12), the variables of firm level characteristics are more significant than those in the models of short-term period. It can be concluded that the risk-taking behavior of privatized banks would be affected by their own firm level characteristics (efficiency, leverage, and profitability) more in the long run.
35
5. Conclusions
Most scholars document the enhanced post privatization performance by privatized firms, but only little evidence is provided to address the issue of risk following privatization (Otchere, 2011). Although the enhanced post privatization performance is important, the risk-taking behavior of state-owned enterprises after privatization also needs to be taken into account, especially when the government plans to privatize the state-owned banks. Because the state-owned banks play such a vital role in the society, they are expected to promote the economy growth and help the stability of the financial system. Thus, my paper is developed to focus on the risk-taking behavior of privatized banks following privatization.
This paper is aimed to examine the risk-taking behavior of privatized banks pre- and post-privatization. The short-term effect and the long-term trend of privatization are investigated in this study, and more variables of country and firm level characteristics are employed as well. In this paper, I provide an empirical analysis on the risk-taking behavior of banks in post privatization. The panel data encompassing both the pre and post privatization periods for 86 privatized banks from 37 countries allows me to examine this issue.
36 privatization. (iv) The influence of the number of privatization transactions is not significant in the long run. (v) The country level characteristics (Devloping, Law, and GDP per capita) and the firm level characteristics (COI, leverage, and ROE) most significantly affect the risk-taking behavior. (iv) The risk-taking behavior of privatized banks is affected by direct privation effect significantly in the short-term (3 years), while the country and the firm level characteristics have more influence on the risk in the long run.
To sum up, I find that the risk-taking behavior of privatized banks is decreasing after privatization, and the foreign ownership contributes to the decrease of risk-taking most significantly. Most of the findings are consistent with the existing literature, such as GDP per capita, leverage and ROE. In the long run, the country and the firm level characteristics affect the risk-taking behavior more significantly than the privatization does.
This study contributes to the existing literature on privatization and risk-taking. It provides a whole picture that how the privatization affects the risk-taking behavior of privatized banks in the short run and in the long run, and it also sheds light on what are the effects of the country and the firm level characteristics on risk-taking.
A better understanding of the changes in bank risk-taking behavior following privatization enables banks to formulate appropriate strategies. According to the regression results, the country and the firm level characteristics matter in explaining the risk-taking behavior of privatized banks. For example, the lower leverage and the higher ROE can benefit the reduction of bank risk. This study also provides the government or policy makers an overview of privatization effects. For example, the share issue privatization and foreign ownership are negatively related to the risk-taking in the long run. So, when the government decides to launch the privatization, the implementation of privatization needs to be fully considered.
37 6.
References
Andrews, A. M. (2005). State-Owned Banks,Stability, Privatization, and
Growth:Practical Policy Decisions in a World Without Empirical Proof. IMF
Working Papers, 05/10.
Beck, T., Cull, R., & Jerom, A. (2005). Bank Privatization and Performance
Empirical Evidence from Nigeria. Journal of Banking and Finance, 29, 2355-2380.
Bonin, J. H. (2005). Privatization matters: Bank performance in transition countries.
Journal of Banking and Finance, 29, 2155-2178.
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.
Boubakri, N., Crosset, J.-C., Guedhami, O., & Omran, M. (2007). Foreign investor participation in privatizations: does the institutional environment matter?
Journal of Financial Research, 30, 129-146.
Chen, Z., Li, D., & Moshirian, F. (2005). China’s financial services industry: The intra-industry effects of privatization of the Bank of China Hong Kong.
Journal of Banking and Finance, 29, 2291-2324.
Clarke, G. C. (2001). Does foreign bank penetration reduce access to credit in
developing countries? Washington: The World Bank, Development Research
Group, Finance.
Dermine, J. (1986). Deposit rates, credit rates and bank capital. Journal of Banking
and Finance , 10, 99-114.
Frank, S. (2013). Bank Insolvency Risk and Z-score Measures: A refinement.
Working papers series.
Fraser, D., & Anderson, R. (2000). Corporate control, bank risk-taking, and the health of the banking industry. Journal of Banking and Finance, 24, 1383-1398. George R.G. Clarke, R. C. (2005). Bank privatization in developing countries: A
summary of lessons and findings. Journal of Banking & Finance , 29,1905– 1930.
38 Gonzalez, F. (2005). Bank regulation and risk-taking incentives: an international
comparison of bank risk. Journal of Banking and Finance, 1153-1184.
Gorton, G. R. (1995). Corporate control, portfolio choices, and the decline of banking.
Journal of Finance, 50, 1323-1354.
Greenbaum, S., Thakor, A., & Chan, Y.-s. (1986). Information reusability competition and bank asset quality. Journal of Banking and Finance, 10, 242-253.
Hellmann, T. F. (2000). Liberalization, Moral Hazard in Banking, and Prudential Regulation: Are Capital Requirements Enough? American Economic Review, 90(1): 147-165.
James R. Barth, G. C. (2002). Bank Regulation and Supervision: What Works Best?
NBER Working Paper Series.
Jia. (2009). The effect of ownership on the prudential behavior of banks- the case of China. Journal of Banking and Finance, 33, 77-87.
John H. Boyd, G. D. (2005). The Theory of Bank Risk-taking and Competition Revisited. The Journal of Finance, 60,1329-1343.
Keeley, M. C. (1990). Deposit insurance, risk, and market power in banking. The
American Economic Review, 5, 1183-1200.
Kwan, S., & Eisenbeis, R. A. (1997). Bank Risk, Capitalization, and Operating Efficiencty. Journal of Financial Services Research, 12, 117-131.
La Porta, R., Lo´pez-de-Silanes, F., & Shleifer, A. (2002). Government ownership of banks. Journal of Finance, 57, 265–301.
Laeven, L. a. (2009). Bank governance, regulation and risk taking. Journal of
Financial Econnomics, 93, 259-275.
Lev, B. (1974 ). On the association between operating leverage and risk. Journal of
Financial andQuantitative Analysis, 9, 627-641.
Lin, X., & Zhang, Y. (2009). Bank ownership reform and bank performance in China.
Journal of Banking and Finance, 1, 20-29.
Marcus, A. (1984). Deregulation and bank financial policy. Journal of Banking and
39 Megginson, W. (2005). The economics of bank privatization. Journal of Banking and
Finance, 29, 1931–1980.
Mohsni , S., & Otchere , I. (2011). BANK RISK-TAKING BEHAVIOR
FOLLOWING PRIVATIZATION. The 19th Conference on the Theories and
Practices of Securities andFinancial Markets (pp. 1-32). Taiwan: NYSU.
Mullins, H. M. (1993). Risk-taking, managerial compensation and ownership structure: An empirical analysis. Working Paper, University of British Columbia.
Omran, M. (2004). Performance consequences of privatizing Egyptian state-owned enterprises:The effect of postprivatization ownership structure on firm performance. Multinational Finance Journal, 8, 73-114.
Otchere, I. (2005). Do privatized banks in middle- and low-income countries perform better than rival banks? An intra-industry analysis of bank privatization.
Journal of Banking and Finance.
Otchere, I. (2007). Does the response of competitors to privatization announcements reflect competitive or industry-wide information effects? International evidence. Journal of Empirical Finance, 4, 523-545.
Otchere, I. (2009). Competitive and value effects of bank privatization in developed countries. Journal of Banking and Finance , 33, 2373-2385.
Otchere, I. C. (2003). Intra-industry effects of bank privatization: A clinical analysis of the privatization of the commonwealth bank of Australia. Journal of
Banking and Finance, 27,949-975.
Rachdi, H. (2011). Board characteristics, performance and risk taking behavior in Tunisian Banks. International Journal of Business and Management, 6, 88-97. Rokhim, R., & Susanto, A. (2011). The Increase of Foreign Ownership and its Impact
to the Performance, Competition & Risk in Indonesian Banking Industry.
Working Paper.
Roy, A. D. (1952). Safety first and the holding of assets. Econometrica, 20, 431-449. Verbrugge, J. M. (1999). State ownership and the financial performance ofprivatized
