Do Privatized Banks in Middle- and Low-Income Countries Perform Better than Rival
Banks? An Intra-Industry Analysis of Bank Privatization
Isaac Otchere*
Department of Finance
Faculty of Economics and Commerce
The University of Melbourne
Parkville, Victoria 3010
Australia
*Email: [email protected] Tel: (61) 3 8344-7166; Fax:(61) 3 8344-6914
1
Do Privatized Banks in Middle- and Low-Income Countries Perform Better than Rival
Banks? An Intra-Industry Analysis of Bank Privatization
Abstract
This paper presents a comprehensive analysis of the pre- and post-privatization operating performance and stock
market performance of privatized banks and rival banks in middle- and low-income countries. First, we find that
privatization announcements elicit negative abnormal returns for rival banks and that the effects are more
pronounced for subsequent tranche sale where the proportion of government ownership in the privatized bank is
reduced. The rival banks’ reaction to privatization announcement is consistent with the predictions of the
competitive effect hypothesis. Second, the privatized banks underperformed the benchmark index in the long run.
Investors who bought shares of the privatized banks on the first day of trading and held them for 5 years (instead
of investing in the market index) lost 25% of their wealth. The underperformance is consistent with the negative
long run returns that have been documented for initial public offerings. Third, we do not find any statistically
significant improvements in the privatized banks’ post privatization operating performance. On the contrary, we
find that the privatized banks carry higher problem loans than their rivals in the post privatization period. In
addition, they appear to be overstaffed relative to their rivals. Our cross sectional regression results indicate that
the proportion of government ownership significantly explains the underperformance of the privatized banks. The
results support the view that continued government ownership of the privatized banks may hinder managers’
ability to restructure the privatized bank.
Keywords: Bank; privatization; rivals’ reaction; returns, operating performance, middle/ low income countries
JEL Classification: G21, G32, G 14
The author thanks Ben Amoako-Adu, Kevin Davis, Paul Kofman, Christine Brown, John Handley, Krishnan
Maheswaran and workshop participants at the Department of Finance, University of Melbourne for their helpful
comments. Financial support provided by the Melbourne University Research Grant Scheme is acknowledged.
Jeffri Gani, Lai Kwan, Kim Choo and Valmik Nagri provided research assistance. All errors remain my
responsibility.
2
Do Privatized Banks in Middle- and Low-Income Countries Perform Better than Rival
Banks? An Intra-Industry Analysis of Bank Privatization
1
Introduction
Prior studies have found that privatization of state-owned enterprises (SOEs) improve the
firms’ performance. Megginson et al., (1994) and Boubakri and Cosset (1998) document strong
performance improvements for their sample of privatized firms in developed and developing
countries respectively. Unlike privatization of non-financial firms where a reasonably large
number of research exists, empirical work on bank privatization is beginning to evolve. Also,
unanswered in previous research is whether there is significant information transfer effect of
privatization on rival firms.
An information transfer effect is the change in the value of a firm that can be attributed
to firm specific announcements made by other firms. We argue that privatization can have
positive information effects or competitive effects on rival firms. Under the competitive effect
hypothesis, the rival firms would react negatively to privatization announcement if the market
believes that there is now a more efficient, aggressive and rejuvenated competitor in the
industry whose operations can lead to falls in product prices and, hence, erode the profitability
of the rival banks. On the other hand, under the information effect hypothesis, privatization
announcement could signal positive information about the industry rivals. In this paper we
examine rival banks’ reaction to privatization announcement and then analyze the pre and post
privatization operating performance and post privatization stock market performance of
privatized banks relative to that of rival banks in middle- and low-income countries.
We study bank privatization in middle- and low-income countries for a number of
reasons. First, like most state-owned firms, government owned banks are characterized by
3
inefficiencies and profitability problems, including low quality loan portfolios that result from
the granting of loans to poorly performing state enterprise and political supporters (La Porta et
al., 2002). However, because of the pivotal role state banks play in the development of an
efficient financial system in middle- and low-income countries, their privatization usually
generates serious opposition from interest groups and the community at large. Second, Perotti
and Guney (1993) argue that banks in emerging economies have strong but perverse incentive
to continue to fund former debtors (i.e., state enterprises) that are less efficient and more risky
than private firms because doing so enables them to gain the potential of repayment of previous
debt granted to them when the bank was a state bank. Given this incentive to continue to fund
risky clients and the usually stiff opposition to bank privatization in these countries, it is worth
examining whether the touted efficiency and profitability gains materialize after privatization.
Third, bank privatization offerings tend to be of substantial size (Verbrugge et al., (1999)).
From the perspective of this study, this is an important feature because such large privatization
offerings are likely to have significant intra-industry effects on rival banks.
Fourth, state banks in middle- and low-income countries provided project finance to
state-owned enterprises that were once the commanding heights of the economy. But, with the
role of state enterprises in most countries significantly curtailed and the SOEs subsequently
sold partly because of their inefficiencies, so too have the state banks that provided the
wherewithal for these SOEs. However, unlike privatization of non-financial firms where a
reasonably large number of research exists (see Megginson and Netter 1998 for a survey of this
literature), very little empirical research exists on bank privatization. The exception to this is a
number of country specific studies (including Clarke and Cull (2002), Brock (2000),
Abarbanell and Bonin (1997) and Otchere and Chan (2003)) and perhaps the only cross-
4
country study by Verbrugge et al., (1999). The country specific studies provide narrow insights
into the effects of bank privatization. The cross-country study, like the non-bank privatization
studies, focuses primarily on the pre- and post-privatization operating performance of
privatized banks. However, such trend analysis has the tendency to obscure true performance
changes if the industry is undergoing changes. Therefore, it is important to ascertain how the
privatized banks’ performance measures up to that of the rival banks in the country. Verbrugge
et al., (1999) argue that this type of comparison allows for the separation of the effects of
privatization from general banking trends. We examine the performance of privatized banks
relative to that of rival banks using both accounting and stock market data for the analyses.
Four analyses are carried out. First, we analyze the short-term stock price reaction of
rival banks around the time of the privatization announcement with a view to inferring from the
rivals’ reaction investors’ expectation of the efficiency of the privatized firms.1 This is followed
by an analysis of the long-term stock market performance of the privatized banks and their
rivals. A comparative analysis of the operating performance of the privatized banks and their
rivals based on a variant of the CAMEL criteria over the pre- and post-privatization period is
carried out.2 Finally, we run cross-sectional regressions aimed at identifying the determinants of
the long-run stock market performance of both the privatized banks and their rivals.
We find that on average, the rival banks’ stock price reacted negatively to news of bank
privatization in middle and income countries, albeit the reaction occurred with a lag. The
shareholders of the rival banks lost 1.32% of their wealth on the announcement day and the
following day. The results suggest that competitive effects outweigh any positive information
1
Eckel et al., (1997) argue that the market’s exp ectation of the efficiency of the privatized firm can be inferred
from the rival firms’ price effects following the privatization announcement.
2
CAMEL is an acronym for Capital adequacy, Asset quality, Management efficiency, Earnings ability and Labor.
5
that the privatization might have conveyed about the industry rivals. Contrary to the findings of
prior studies, privatized banks in middle and low income countries do not experience any
significant improvements in operating performance in the post privatization period. Rather, we
find that on a relative basis, the privatized banks carry higher loan loss provisions and higher
problem loans. In addition, they appear to be over-staffed relative to their rivals. The lackluster
operating performance of the privatized firms reflected in their stock market performance as
they underperformed the benchmark index in the long run. We find that the opportunity cost of
investing in the privatized bank is about 25%. That is, a $1 invested in the shares of the
privatized banks by investors who bought the share on the first day of trading and held the
stock for five years would be worth $0.25 less had it been invested in the market index. The
long run stock market performance of our privatized bank sample is consistent with the long
run underperformance that has been documented for initial public offerings. The long run stock
market performance of our sample is significantly related to operating performance measures,
thus suggesting that the fall in stock price is real loss. This study adds a useful perspective to
previous studies and provides additional information on the implication of privatization
announcements on other firms in the industry. Our results, especially the short run results,
would be relevant to investors and risk arbitrageurs, as they suggest that profits can be made
from trading the shares of not only the privatized banks but also the shares of the rival banks.
The remainder of the paper is organized as follows: Section 2 presents a discussion of
the political economy of bank privatization. Our testable hypotheses regarding the effects of
privatization on the share price of rival banks are presented in Section 3. Section 4 describes
the data. In Section 5, we present evidence on the stock market reaction of rival banks to the
privatization announcement and the long run stock market performance of the sample. Section
6
6 considers evidence on the operating performance of the sample while cross-sectional
regression results are presented in Section 7. A summary is provided in Section 8.
2.
Political economy of bank privatization
The role of state banks in economic development has been the subject of a number of studies.
Generally, there are two broad views regarding government ownership of banks, namely, the
development and political views.3 The development view asserts that in countries where the
financial system is not well developed for private banks to play the crucial developmental role
of channeling capital into productive sectors, governments step in, through their ownership of
banks, to perform the role of development banks by collecting savings and directing them
towards strategic, long term projects (La Porta et al. 2002). The political view sees government
ownership of banks as a tool to achieve political objectives through the provision of
employment and subsidies to supporters in return for political contributions and votes (Shleifer
and Vishny, 1994). This view suggests that government ownership of banks and the resulting
politicization of resource allocation would slow down financial development. La Poter et al
(2002) provide empirical evidence to support the view that government ownership of banks
makes the financial system less efficient.
In the past few decades, middle- and low-income countries had near-monopoly banking
system and state banks, acting as development banks, provided project finance to SOEs that
were once the commanding heights of the economy. As it turned out, the projects that were
funded by government finances were inefficient, perhaps because of the conflicting economic,
social and political goals. Consequently, a large number of state enterprises have been
3
See La Porta et al., (2002) for a detailed review of these schools of thought.
7
privatized. With the role of state-owned enterprises in most countries significantly curtailed and
the enterprises restructured or privatized, so too have the banks that provided the finances for
these SOEs. The reforms in the middle- and low-income countries, which are sometimes
recommended and financially supported by the IMF and the World Bank as part of Structural
Adjustment Program (SAP) in the last two decades, have made ownership of banks by
governments less attractive. Bank privatization in middle and low-income countries is thus
considered as a major step in the painful process of disengaging the state from complete control
of the economy.
Given the significant role of state-owned banks in resource allocation/misallocation
however, their privatization becomes a challenging task for governments, especially those in
developing countries where privatization occurs as part of the transition to a more market-based
economic system. From the perspective of the political view, bank privatization means the
beneficiaries of government misallocation of resources will lose their spoils, including
subsidized loan, employment and other benefits. As a result, bank privatization generates
serious opposition from interest groups and the community at large. In addition economic
restructuring in these economies generates a high degree of political instability as a result of the
uncertainties associated with the transition from state-apparatus to market based economies,
thus making bank privatization even more difficult under those circumstances. As Verbrugge et
al (1999) argue, opposition from entrenched parties makes bank privatization as important a
tool of political economy as it is a mechanism for banking reform. Despite the serious
opposition that usually accompanies bank privatization however, a significantly large number
of banks from developing countries have been privatized (at least partially) during the past two
decades.
8
Although the role of state enterprises in economic development in middle- and lowincome countries has decreased and the benefits associated with owning a poorly performing
bank has become less attractive than before, government ownership of banks in developing
countries is widespread since credit and direct grants from the government are considered
important lubricants as SOEs still play a significant role in economic development in these
countries.4 Under these circumstances, governments feel reluctant to fully privatize state bank.
Berglof and Roland (1998) argue that the unwillingness of governments to allow troubled
SOEs to go bankrupt and/or workers to be laid off as a result of the restructuring suggest how
painful true banking reforms could be. Hence, political pressure or economic reasons belie the
partial privatization of state banks and not even some prodding from the SAP sponsors have led
to large scale full privatization of banks in middle- and low-income countries.
Since the relationship with the government is not completely severed, partially
privatized banks may continue to subsidize former debtors by granting them concessionary
loans. Perotti and Guney (1993) show that banks in certain emerging economies have strong,
but perverse, incentive to continue to fund former state owned enterprises although these
enterprises are more risky than private firms. By doing so, they gain the potential of repayment
of previous debt granted to them when the bank was a government-owned bank. This suggests
that debt overhang could affect the performance of the privatized banks; hence examining the
performance of privatized banks relative to that of rival banks with a view to ascertaining
whether or not performance improves after privatization is interesting in its own right.
4
The use of banks as vehicles to promote economic development is not restricted to only emerging economies.
Verbrugge et al., (1999) argue that even in market-oriented economies like the US, banks and credit allocation
schemes are favored vehicles for enhancing credit availability to favored sectors or groups.
9
3.
Effects of privatization on competitors and testable predictions
3.1
Competitive effects
Privatization usually brings about a change in the firms’ objective and this, together with a
change in the owners and managers’ incentive, often leads to a more focused and efficient
organization (Megginson et al., (1994) and D’Souza et al., (2000)). The public trading of the
firms’
shares
also
facilitates
the
adoption
of
market-oriented compensation plans, as
management compensation can be tied to the firm’s stock price. This creates an incentive for
management to perform better. The pressures of product market competition may also compel
the newly privatized firms to operate more efficiently, aggressively and competitively if they
are to survive in the post privatization period. Furthermore, following privatization the firms
could retain significant market power (on account of their hitherto monopoly status) while
being relieved of the requirements to follow government directives designed to promote social
aims. The privatized firms may be able to exploit this market power to their advantage to
increase profitability. The former SOE could thus become a stronger competitive force to the
rivals in the industry because of their dominant position in the product market. Thus, the
privatization of a firm could hurt rivals through increased competition.
From this perspective, privatization announcement could signal information about
increased level of competition in the industry. If the market believes that there is now a
rejuvenated, more efficient and aggressive competitor in the industry whose operations can lead
to falls in product prices and, hence, erode the profitability of the rivals, then the rival firms’
stock prices will react negatively to the privatization announcement.
10
3. 2
Positive information effects
Privatization announcements could also generate positive information effects for the rival firms
in a number of ways. First, privatization could lead to the relaxation of rules of operations in
the industry. Deregulation that accompanies or precedes privatization could unlock growth
opportunities for all firms in the industry. Second, the presence of the newly privatized, often
large and visible firm in the industry could attract more analyst coverage and institutional
interest in the industry. This increased attention and focus can lead to efficient valuation of the
industry as a whole (Subrahmanyam and Titman (1999)). The benefit is similar to investment
bankers’ claim that floating equity in businesses not previously exposed to the market makes
their operating performance more transparent and raises shareholder returns by revealing
hidden value. To the extent that investors perceive that rival firms will benefit from
improvement in the quality of analysts’ coverage of the industry as a whole, the rival firms’
stock price could react positively to privatization announcements.
Third, prior studies show that privatization leads to improved performance for the
privatized firm. We argue that the presence of a rejuvenated competitor in the industry could
spur the rival firms to perform better if they are not to lose market share to the privatized firm.
Hence, if the market believes that rival firms could achieve parallel gains as the privatized
firms, then their stock price would increase in response to privatization announcement. This is a
‘shot-in-the-arm’ effect. Fourth, privatization usually results in the loss of non-competitive
incentives such as subsidies and tax cuts that the former SOE used to enjoy. The loss of these
subsidies could hurt the competitive position of the newly privatized firm relative to that of the
rivals. Alternatively, the fact that all firms are now operating on an equal playing field will
make the rival firms more competitive than before. The foregoing discussion suggests that
11
privatization announcements could convey positive information about the future prospects of
the industry rivals.5
4.
Data
The list of privatized banks was identified from Verbrugge et al., (1999) and the supplemental
appendix to Megginson (2000). We use the World Bank country classification list to identify
the privatized banks that fall into middle- and low-income countries category. The privatization
announcement dates were identified from Reuters business news archives.6 All publicly traded
banks that existed in the country at the time of the privatization were initially considered as
rival banks. The stock price and financial statement data come from Datastream International
and Bank Scope databases respectively. To be included in the study, the privatized bank should
have rivals in the country at the time of the privatization announcement. Also, we require that
the privatization announcement date and stock price and financial statement data be available in
the aforementioned sources. To reduce confounding effects we exclude rival banks that
announced any significant event such as mergers or earnings announcement around the event
period. For 18 banks that were privatized between 1989 and 1997 and 28 rival banks, we were
able to obtain the necessary data. For three of the privatized banks, the respective governments
(from India, Philippines and Poland) sold additional shares during the study period. Thus in
total, 21 privatization observations are analyzed in this study.
[Fix Table 1 here]
5
Prior studies have shown that investors use announcements by one firm to make inferences about other firms that
operate in the same industry. Corporate announcements for which information transfer effects have been
documented include bank failures (Aharony and Swary, 1983), merger proposals (Eckbo, 1983), takeovers
(Akhigbe and Madura, 1999), bankruptcy announcements (Lang and Stultz, 1992), going private events (Slovin,
Sushka and Bendeck, 1991), public offerings of common stock, convertible debt and straight debt (Szewczyk
1992), share repurchases (Hertzel, 1991 and Otchere and Ross, 2002)) and dividends (Firth 1996).
6
The announcement date relates to the first time that the government announced its intention to privatize the firm.
12
Descriptive statistics for the sample are presented in Table 1. Panel A shows the distribution of
the sample while Panel B presents summary statistics on the share issue. Governments in nine
countries carried out the privatization transactions. The data show that most of the bank
privatizations occurred in the latter half of the 1990s. The mean (median) gross proceeds from
the privatization are US$156 million (66 million). While about 7% of the issues were reserved
for employees, a third of the shares were issued to foreign investors. A feature of the sample
worth highlighting is that most of the banks (88%) are partial privatized, with the mean
(median) percentage of government ownership after the privatization being 49% (53%). This
finding is consistent with the view that although the benefits of owning a poorly performing
bank has become less attractive than it had been before because of the privatization of SOEs,
yet governments in middle- and low-income countries find it difficult to release control over
state banks.
Privatization usually takes a long time to yield gains as more time may be required by
management of the privatized firm to overcome organizational inertia and resistance to change
that usually characterizes newly privatized firms. In the case of banks, Verbrugge et al., (1999)
submit that it takes substantial time to both develop appropriate market-oriented lending
policies and to accumulate properly priced loan portfolios. Therefore, to more broadly
investigate the effects of privatization on the sample firms, we use a much longer study period
than has generally been used in private sector IPO studies. Specifically, we use 3 years pre
privatization operating performance data and 5 years post privatization operating performance
and stock market data to assess the performance of the sample.7
7
Since most of the privatization occurred in the latter half of the 1990s (see Table 1), a much longer (post
privatization) window would significantly reduce our sample of privatized banks because of data limitations.
13
5.
Stock price effects of bank privatization
5.1
Rival banks’ reaction to bank privatization announcements
5.1.1 Design
The rival banks’ reaction to privatization announcements is calculated on a market–adjusted
basis.8 To obtain market-adjusted daily abnormal returns, we adjust individual bank’s daily
returns for the contemporaneous return of the market. Designating the announcement date as
day 0, we then estimated the abnormal returns over the 21 days surrounding the announcement
date. The daily market-adjusted abnormal returns were then averaged across all banks.
Cumulative abnormal returns were calculated by summing the daily market-adjusted returns
across different event windows. Then t-tests were conducted by dividing the abnormal returns
by their contemporaneous cross sectional standard errors.
5.1.2
Results
The abnormal returns realized by the rival banks at the time of the privatization announcements
are presented in Table 2. Consistent with the assertion that privatization could hurt rivals, we
find that the rival banks reacted negatively to the privatization announcements. While the
abnormal returns for all the event windows are negative, there appears to be a delayed reaction
to the privatization announcements, as statistically significant negative abnormal returns are
documented only for the period following the announcement date. The shareholders of the rival
banks lost 1.32% of their wealth on the announcement day and the following day. By the end of
the third day following the privatization announcement, the rival banks had lost 2.80% of their
wealth (t-statistic = 2.49), with 63% of the rival banks exhibiting negative abnormal returns.
8
Since the privatized banks did not have stock price data, we use the market-adjusted method to compute their
long run returns. To be consistent, we use the same method to compute the rival banks’ short run abnormal returns
following the privatization announcement.
14
The reaction seems to have lasted for one week, as the cumulative abnormal returns
documented thereafter are not statistically significant at conventional levels. Given the small
sample size (for which we were able to obtain daily data), we also report the median abnormal
returns for the rival banks. While the median returns are relatively smaller in magnitude, they
are nonetheless all negative, with the median abnormal returns of –1%, –1.33% and –1,72% for
the [0,3], [0,4] and [0,5] event windows being significant at 10%.
[Fix Table 2 here]
A noticeable feature of our sample of privatized banks is that most of them are partially
privatized. We argue that the rival firms’ reaction to privatization announcement could depend
on whether or not the privatization is part of a continuing program of sell off. If investors
perceive that the government is embarking on a privatization program that entails the gradual
sale of the state bank, then the valuation effects of the ‘initial’ privatization announcement and
subsequent privatization announcement on the rival firms’ stock price could be different. For
privatizations that occur in tranches, the market may learn from the initial privatization
announcement and since investors would expect subsequent privatization to follow, the first
privatization announcement is likely to contain more surprise. Therefore the initial privatization
announcement could elicit stronger reaction from rivals than subsequent partial privatization
announcements.
Alternatively, if government control bodes ill for the SOE’s performance, then the
continued
government
ownership
may
weaken
any
competitive effects of privatization
announcement on the rival banks. Hence, the valuation effects of the initial partial privatization
(IPO) on privatized banks and the rivals could be less than that of subsequent tranche sale that
15
reduces government ownership of the state bank. As the proportion of government ownership
reduces in subsequent partial privatization announcements and the firm gradually moves
towards full privatization, the privatized bank may become more efficient and competitive and
consequently,
the
rival
firms
could
react
more
strongly
to
subsequent
privatization
announcement than the ‘initial’ partial privatization announcement. To test these conjectures,
we examine the rival banks’ reaction to the first and subsequent privatization announcements.
The results presented in Panel B of Table 2 generally support the latter conjecture. We find that
while the rival banks’ reaction to the initial partial privatization announcement is not
significant,
their
reaction
to
subsequent
privatization
announcements
is
significant.
Shareholders of rival banks lost over 2% in the days following the announcement of further
sale of government shares in the privatized banks. The rival banks’ reaction to subsequent
privatization announcement seems to drive the combined results presented in Panel A.9 The
results suggest that as the proportion of government ownership in the privatized banks reduces,
subsequent partial privatization announcements generate stronger adverse stock market reaction
from rival banks.
In summary, we note that the consistently negative stock price response of the rival
banks to privatization announcements provides evidence that is broadly consistent with the
competitive effects hypothesis and suggests that investors view privatization announcements as
foreshadowing bad news for rival banks. Our results parallel previous evidence that shows that
a number of corporate events, including going private transactions (Slovin et al 1991),
bankruptcy announcements (Lang and Stulz, 1992), security offering (Szewczyk, 1992),
9
While the ‘initial’ privatization announcement may not be the first privatization because the governments had
sold part of the shares earlier (sometimes by trade sale), the results nonetheless indicate that as the percentage of
government ownership reduces, the rival banks’ adverse reaction becomes greater.
16
dividend changes announcement (Firth, 1996 and Laux et al., 1998) and takeovers (Akhigbe
and Madura, 1999), have implications for rival firms.
5.2
Long run stock returns to investors in bank share issue privatization
5.2.1 Design
The long run stock market performance of the sample firms is examined by analyzing the
returns of the privatized banks and those of the rival banks. We hypothesize that privatization
would give the management of the privatized banks the liberty to pursue growth-oriented (but
perhaps politically unpalatable) policies that will enable the banks to generate higher returns for
investors. To examine this conjecture, we compare the 5-year post privatization returns of our
privatized banks sample with the returns of rival banks. Since the privatized banks did not have
stock market data prior to the initial privatization date, we use the market-adjusted and buy and
hold methods to compute abnormal returns.
The market-adjusted abnormal return is the difference between the individual bank’s
returns and returns of the benchmark index in the respective country. The abnormal returns are
calculated using the same procedure as described in section 5.1 except that here, we use
monthly returns. The buy and hold returns are the returns realized for buying the shares on the
first day of trading and holding them for a period of 12, 24, 26, 48 and 60 months. Following
Byun and Rozeff (2003), we calculate the buy and hold abnormal returns as:
n
n
t =1
t =1
BHARit = ∏ (1 + Rit ) − ∏ (1 + Rmt )
(1)
where BHAR is the market-adjusted buy and hold abnormal returns, Ri is the returns for bank i,
RM is the returns on the market index and n is the end of the holding period. The abnormal
returns for each month are obtained by taking the average across the sample. These returns are
17
subsequently cumulated over different periods over the 5 years following the privatization. We
conduct t-tests by dividing the abnormal returns by their contemporaneous cross sectional
standard errors.
5.2.2 Results
The returns using the 2 methods are summarized in Table 3. The market-adjusted abnormal
returns for the privatized banks and those of the rivals are presented in Panel A, while the buy
and hold returns are presented in Panel B. The results show that the privatized banks
underperformed the market in the years following the privatization. The privatized banks’
cumulative market-adjusted abnormal returns for the fifth year of privatization are –30% and is
significant at 10%, though these returns are not statistically different from those of the rival
banks. The percentage of privatized banks that earned positive abnormal returns increased from
36% at the end of year one to 53% at the end of year three, but subsequently reduced to only
33% at the end of year 5, while about half of the rival banks earned positive abnormal returns
during the period.
[Fix table 3 here]
Similar results are obtained when returns are calculated using BHAR method. Investors who
bought the shares of the privatized banks on the first day of trading and held the stock for 5
years (instead of investing in the market index) lost 24% of their wealth. Only about 25% of the
privatized banks earned positive abnormal returns over the 5 years following the privatization.
The median privatized bank also underperformed the median rival bank at the end of year 5,
with the difference in performance of –20.75% being significant at 1%. The month-by-month
market-adjusted cumulative abnormal returns of the privatized banks and those of the rivals are
18
presented graphically in figure 1. We note that the privatized banks outperformed the market
and the rivals in the four months following the privatization. The improved performance in the
period
immediately
following
the
share
issue
perhaps
reflects
the
well-documented
underpricing of initial public offerings. For government IPOs, the underpricing could be
widespread and substantial since that would ensure that shares in subsequent government
privatizations are well subscribed.10 The privatized banks began to underperform their rivals
almost six months after they had been privatized. Thereafter, their performance has been on the
slide while the rivals’ returns have been relatively stable, although slightly lower than that of
the benchmark index.
In summary, both the market-adjusted and buy-and-hold returns show that the
privatized banks underperformed the market, and to a small extent the rival banks. Our long run
results contrast with those of Megginson et al., (2000), Levis (1993) and Menya and Paudyal
(1996) who find that privatized firms earn positive abnormal returns in the long run. Our results
are however consistent with 2 strands of literature. First, they are consistent with the results of
several studies that document significantly negative long run returns to initial public offerings
(see for example, Ritter, 1991, and Loughran et al, 1994). Second, the results are consistent
with those of Otchere and Chan (2003) who find that a partially privatized Commonwealth
Bank of Australia underperformed the market. As shown in Table 1, about 88% of the banks
are partially privatized so the underperformance may reflect the difficulties inherent in
improving the performance of a partially privatized bank whose major shareholder is the
government.
10
Underpricing in government IPO could also be significant because unlike private sector IPO where the
overriding consideration influencing pricing is maximization of proceeds, governments may use privatization to
pacify employees who might hinder the privatization program or to build domestic political support (Dewenter and
Malatesta (1997).
19
6
Operating performance changes
6.1
Design
We examine the pre- and post-privatization operating performance of the privatized banks
relative to that of the rival banks using the CAMEL criteria. CAMEL, as used in this study,
stands for Capital adequacy, Asset quality, Management efficiency, Earnings ability and Labor
(employment levels and productivity). The Federal Deposit Insurance Corporation and other
researchers including Thomson (1991) and Persons (1999), have used the CAMEL criteria to
assess the performance of banks. We analyze 12 ratios grouped under the CAMEL criteria.
Details of the ratios are listed in Appendix 1.
(a)
Capital Adequacy
The capital adequacy ratio is the total of Tier 1 and Tier 2 capital and is measured as the ratio
of capital to risk-adjusted assets and off-balance sheet exposure determined on a risk-weighted
basis of at least 8%. A higher ratio reflects the bank’s ability to absorb unanticipated capital
losses.
(b)
Asset Quality
This criterion relates to the impairment of bank loans, the asset with the highest probability of
deterioration. We analyze three loan-related measures that the banking industry uses to measure
loan quality problems namely, provisions-to-total loans, the ratio of impaired assets (nonperforming loans) to total loans, and net impaired assets-to-total loans. The degree to which
provisions are made in anticipation of, or concurrent with, actual impairment in the loan
portfolio reflects credit quality. A privatized bank may aggressively build its loan portfolio and
could be forced to make large provisions for unanticipated bad debts. It is also possible that a
privatized bank may be more efficient in managing its loan portfolio and therefore carry only a
20
small loan loss provision. We analyze provisions-to-loans ratio to ascertain how efficient
privatized banks have managed their loan portfolios.
We recognize, however, that banks can smooth incomes by making higher provisions
than necessary when credit quality and net income are high. Consequently, they may not
increase provisions as much if credit quality is deteriorating. Gunther and Moore, (2000) argue
that income smoothing will ensure that banks with asset quality problems can raise net income
and retained earnings, thereby boosting Tier 1 capital. As a result of the limitations of
provisions-to-loans ratio as a measure of asset quality, we use the ratio of gross impaired
assets-to-loans and net impaired assets-to-loans as additional measures of asset quality, as these
ratios alleviate the problem that banks may have underestimated their loan provisions. Higher
ratios reflect poor asset quality.
(c)
Management Efficiency
We employ operating efficiency measures such as cost-to-income (operating expense to
operating income) ratio and expense-to-asset ratio as proxies for management efficiency.
Lower ratios reflect higher management efficiency. We expect that because of greater emphasis
on profitability, the privatized banks would be more efficient than they were prior to the
privatization.
(d)
Earnings ability (profitability)
Privatization leads to the transfer of both control and cash flow rights to managers who are
more interested in profits and efficiency than the politicians. Therefore, we expect the
profitability of the privatized banks to increase following their privatization. We use net interest
margin and return on assets (ROA) as measures of profitability. However, as argued by
21
Rhoades (1998), ROA is baised upwards for banks that earn significant profits from off-balance
sheet operations such as derivative activities, as these activities generate revenue and expenses
but not assets. Consequently, we employ return on equity (ROE) as an alternative measure of
profitability. Higher ratios indicate improvement in performance.
(e)
Labor (growth in staff levels and employee productivity)
There is evidence that governments use banks to provide employment and subsidies to
supporters in return for political contributions and votes. This political interference sometimes
compels the bank to maintain unprofitably business units. To ascertain whether a significant
change in employment and labor productivity occurs after privatization, we analyze three laborrelated ratios. First, we use the ratio of asset-to-number of employees as a proxy for overstaffing and then analyze growth in staff levels to ascertain whether the privatized banks
reduced staff levels after privatization. Finally, we use the ratio of total revenue-to-number of
employees to measure employee productivity. We conjecture that privatized banks will place
more emphasis on profitability and are therefore more likely to reduce employment and
improve employee productivity after privatization.11
The change in performance is first examined by comparing the privatized banks’ ratios
from year –3 to year +5. Analyzing the trend in performance over the pre and post privatization
periods is perhaps not adequate because it may be difficult to draw conclusion from the result,
especially the mean ratios because these data do not adjust for economy-wide or industry-wide
factors that may affect the ratios. As Cornett and Tehranian (1992) argue, any trend in the
11
Managers are also likely to reassess economically unprofitable investments in branch network. The lack of data
precludes an analysis of this variable. However, given the high correlation between bank branches and number of
employees, we analyze growth in staff levels to ascertain whether or not the privatized banks reduced their staff
levels since privatization.
22
banking industry would affect bank ratios. Thus any significant change documented for the
privatized firm could be due to factors other than the privatization. To account for the impact of
possible contemporaneous events we also report industry-adjusted median (mean) performance
measures for the privatized bank. We calculate industry-adjusted performance of the privatized
firms as the difference between the privatized banks’ ratios and the rival banks’ ratios. The
median (mean) ratios of the rival banks thus provide a basis for analyzing the relative
performance of the privatized banks.
The difference in the performance for each year from year –3 to year +5 is tested using
the Wilcoxon signed-rank test calculated as:
z=
w − n (n − 1) / 4
(2)
n ( n − 1)( 2 n − 1) / 24
where z is the Wilcoxon test statistics, w is the sum of the positive ranks, n is the number of
observations,
n( n − 1)
is the mean of w, and
4
n ( n − 1)( 2 n − 1) / 24
is the standard deviation of w.
The significance of the mean change in the pre-privatization period (year -3 to year –1) and the
post-privatization period (year 1 to year 5) performance is examined by performing a t-test.
6.2
Results
The operating performance results are presented in Table 4. Panel A shows the ratios for the
privatized banks while Panel B contains the ratios for the rivals. The results of the Wilcoxon
signed rank test for the difference in median ratios are presented in Panel C. Table 5 also
contains the difference in mean tests for the pre- and post-privatization performance of the
privatized banks and the rivals. The figures presented in Table 4 show that for most of the preprivatization period, the privatized banks were undercapitalized relative to their rivals; or rather
23
the rival banks were overcapitalized relative to the privatized banks. The latter is likely to be
the case since the rival banks’ median ratio decreased in the post privatization period in line
with the privatized banks’ ratio. However, the privatized banks’ mean capital adequate ratio is
greater than that of the rival banks at 5%.
[Fix Table 4 here]
In terms of asset quality, we find that the privatized banks’ provisions-to-loans ratio has
reduced in the post privatization period. The post privatization ratio of 2.30% is less than the
pre privatization ratio of 3.91% at 5%. Despite this improvement however, the privatized banks
appear to carry higher loan loss provision than the rival banks in the post privatization periods.
The privatized banks’ mean post privatization provisions-to-loan ratio of 2.30% is significantly
different from that of the rivals of 1.35 % at the 1% level.
The provisions ratios should be seen in light of the limitations of income smoothing. Since
banks may smooth incomes by over- or under providing for problem loans, we also analyze the
ratio of impaired assets-to-total loans. We find that although the privatized banks’ provisions
ratio has been on the decline in the post privatization period, they carried much higher nonperforming loans than their peers in the post privatization period. 12 The post privatization mean
gross impaired assets-to-total loans ratio for the privatized banks of 14.76% is significantly
greater than that of the rivals of 9.33% at the 10% level.
[Fix Table 5 here]
The privatized banks have also not been able to reduce cost. Rather, their cost-to-income
ratio has increased in the post privatization period although the mean post privatization ratio is
12
We do not have adequate data to analyze problem loans ratio in the pre -privatization period, as most of the rival
banks did not report the value of their gross impaired assets.
24
not statistically different from the pre-privatization ratio. The median expense-to-assets ratio
for the privatized banks has reduced from 16.36% three years before privatization to 12.94%
five years after the privatization, albeit the mean post privatization ratio of 14.65% is not
significantly different from the pre privatization ratio of 12.25%. However, we find that the
privatized banks have been less efficient than their rivals in reducing expenses. While the
privatized banks had similar median expense ratio as that of the rivals in the pre-privatization
period, their expense ratio had become worse in the post privatization period. The rival banks’
expense-to-asset ratio of 10.36% in the post privatization period is significantly lower than that
of the privatized banks of 12.25% at the 5% level.
In terms of earnings ability, there does not appear to be any significant change in the
performance of the privatized banks. Consistent with most of the other operating performance
measures, the mean pre privatization ROE, ROA and net interest margin ratios are not different
from the post privatization ratios, neither are they different from those of the rival banks. We
also find that the privatized bank did not layoff employees; rather they reduced the rate of
growth in employment after privatization. The privatized banks’ mean growth in staff level in
the post privatization period is not statistically different from that of the rival banks. However,
the privatized banks still appear to be overstaffed relative to their rivals in the post privatization
period. Their mean post privatization asset-to-employee ratio (a proxy for overstaffing) of
US$0.66 million per employee is significantly lower than that of the rival banks of US$ 0.93
million at the 5% level.
Two points from our operating performance results are worth emphasizing. First, the trend
analysis
indicates
that
there
is
no
statistically
significant
improvement
in
operating
performance for the privatized banks in the pre- and post-privatization period (apart from a
25
reduction in the provisions ratio). Second, on a relative basis, we note that while the
performance of the privatized banks has been at par with that of the rivals on most of the
operating performance measures in the post privatization period, they have underperformed
their industry counterparts on asset quality and cost efficiency. The privatized banks’ problem
loans have increased in the post privatization period. Expense to asset ratio has become worse
than that of the rivals in the post privatization period. Although their growth in staff is lower,
the privatized banks appear to be significantly overstaffed relative to rivals.
Our results are different from those reported by Otchere and Chan (2003) and
Verbrugge et al., (1999). The difference in results could be due to two reasons. First, the poor
performance of our sample could be due to the fact that it is drawn from middle and lowincome countries. Megginson et al., (1994) suggest that capital market monitoring that
accompanies privatization triggers post privatization improvements, but Holstrom and Tirole
(1995) argue that the benefits from capital market monitoring depends on the level of
sophistication of the market and the intensity of the monitoring. A well-developed and active
capital market allows the newly privatized firms greater access to capital needed to finance
profitable projects. For that reason, privatization in developing countries may offer less
efficiency gains and performance improvements for the privatized firms. In addition, the
privatized banks performance may not significantly improve because they may face less capital
market monitoring since the shareholders would have less access to information and also lack
the power to sanction managerial performance.
Second, the underperformance could be due to the fact that the sample firms are mostly
partially privatized firms. While it has been documented that privatization often leads to
improvement in efficiency and profitability, the degree to which these benefits can be realized
26
depends on whether or not the government fully privatizes the enterprise. A noteworthy feature
of our sample is that many of the banks are partially privatized. Continued government
ownership of the privatized banks may hinder the managers’ ability to restructure the banks.
This probably can explain why the privatized banks are overstaffed relative to their rivals. Prior
studies that analyze the effects of partial and full privatization find that fully privatized firms
are more efficient and profitable than mixed enterprises. Eckel et al (1997) for example find
that Air Canada significantly improved its operating efficiency after the airline had been fully
privatized. Brock (2000) also finds that performance improvements for banks
appear to occur
only after connections to the government and the associated soft budget constraints are severed.
7
Determinants of long run stock market performance
We employ a number of variables to examine the determinants of the long run performance of
the sample firms in a multivariate cross sectional analysis using the following regression model
for both the privatized banks and the rivals:
CARi(1,60)=α+β1 Sizei+β2 LowD+β3 %Govi +β4 FullD+β5 Prodi+β6 CTIi+β7 SEO+β8 LLPi+ ε i (3)
where CARi(1,60) is the 5 year-market-adjusted cumulative abnormal returns;
Size is the logarithmic of the mean post-privatization total revenue;13
LowD is an indicator variable equal to one if the firm is from a low-income country and
zero of it is from a middle-income country;
%Gov is the percentage of government ownership of the privatized bank after
privatization;
FullD is the full privatization dummy and it takes on a value of 1 if the transaction is
full privatization and 0 if it is partial privatization;
Prod is the mean post privatization employee productivity variable;
CTI is mean post-privatization mean cost-to-income ratio;
SEOD is an indicator variable equal to one if the transaction is subsequent equity
offering and zero if it is an IPO and
LLP is the mean post privatization loan loss provisions ratio.
13
We use Total revenue as a proxy for size instead of assets because some banks earn significant profits from offbalance sheet operations such as derivative activities that generate revenue but not assets.
27
The results of the cross sectional regression are presented in Table 6. Consistent with
the conjecture that continued government ownership may affect the performance of the
privatized bank, we find that the coefficient of %Gov is negative and significant, suggesting
that the greater the percentage of government ownership, the worse the post privatization
performance. It appears that the privatized banks’ performance depends on whether the
transaction is an IPO or a subsequent tranche sale. The long run stock market performance is
worse for seasoned equity offerings. This finding is contrary to the expectation that subsequent
sale that leads to a reduction in government ownership of the privatized bank will enhance the
performance of the firm. However, this result should be interpreted in light of the fact that most
of the banks are partially privatized. Hence, the benefits of corporate governance resulting from
private ownership will not occur as long as the government owns the majority shares in these
banks. The result is however consistent with the findings of research in the seasoned equity
issue literature, suggesting that as owners, governments issue more shares when the stock price
is overvalued, hence investors revalue the shares of the firms accordingly. We find that the
larger the privatized firm, the better the long run returns. Privatized banks that have higher
employee productivity perform better. Also, the underperformance documented for our sample
is less for fully privatized banks.
[Fix Table 6 here]
Rival banks that have higher cost-to-income ratios perform worse in the post
privatization. Whether or not the privatization is a subsequent sale does not affect the long run
performance of the rival banks. Interestingly however, we find that rivals of banks that have
been fully privatized perform worse. This is consistent with the prediction of the competitive
effect hypothesis that full privatization can enhance the privatized firms competitive position
28
and that in turn can affect the performance of the rivals. The coefficient of LowD also suggests
that rival banks in low-income countries where relatively less is known about the competitive
force of government banks perform worse than rival banks in middle-income countries. The
explanatory power of the regressions is reasonably high, with the adjusted R2 for the privatized
banks and rival banks regression being 88% and 28% respectively. The F-statistic for both
regressions is significant at 5%. The results for the privatized banks should however be
interpreted with caution because of the small sample size.
8
Summary and Conclusion
In this paper, we present a comprehensive analysis of the pre- and post-privatization
performance of privatized banks and that of their rivals in middle and low-income countries.
First, we analyze the short-term stock price reaction of rival banks around the time of the
privatization announcement and find that the rival banks reacted negatively to privatization
announcements. The negative effects are more pronounced in the case of subsequent tranche
sale where the proportion of government ownership in the privatized banks is reduced. The
negative share price response of the rivals to privatization announcement provides evidence
that is consistent with the competitive effects hypothesis and suggests that investors view
privatization announcements as foreshadowing bad news for rival banks.
Second, both the market-adjusted and buy-and-hold returns indicate that the privatized
banks underperformed the market and, to some extent, the rival banks. The opportunity cost of
investing in the privatized banks is 25%. That is, investors who bought the shares of the
privatized banks on the first day of trading and held them for 5 years (instead on investing in
the market index) earned 25% less. Our results are consistent with the results of several studies
29
that document significantly negative long run returns for initial public offerings. Third, apart
from a reduction in loan loss provisions ratio, we do not find any statistically significant
improvement in operating performance for the privatized banks in the pre- and postprivatization period. On relative basis however, we find that the privatized banks carry higher
loan loss provisions, have larger problem loans and appear to be overstaffed relative to their
rivals in the post privatization period. We conclude that the expected improvement in the
performance of the privatized banks has not materialized. The lack of improvement in
performance could be due to the fact that our sample is from middle and low-income countries
where privatized firms may face less capital market monitoring. But more importantly, the
underperformance could be due to the fact that the sample firms are mostly partially privatized
banks. Thus, the continued government ownership of these banks might have hindered the
managers’ ability to restructure the firms.
30
References
Abarbanell, J.S, and J.P Bonin, 1997, Bank privatization in Poland: The case of Bank Slaski,
Journal of Comparative Economics 25, 1: 31-61
Aggarwal, R, and P. Rivoli, 1990, Fads in the initial public offering market? Financial
Management 19, 45-57
Akhigbe, A. and J. Madura, 1999, ‘The Industry Effects Regarding the Probability of
Takeovers,’ The Financial Review 34, 1-18.
Aharony, J. and I. Swary, 1983, Contagion Effects of Bank Failures: Evidence from Capital
Markets, Journal of Business, 56, 305-22.
Aussenegg, W., 2000, Privatization versus private sector initial public offerings in Poland,
Multinational Finance Journal 4, 1&2, 69-99
Berglof, E, and G. Roland, 1998, Soft budget constraints and banking in transition economies,
Journal of Comparative Economics 26, 18-40
Boardman, A., and A. R. Vining, 1989, Ownership and performance in competitive
environments: A comparison of the performance of private, mixed and state-owned
enterprises, Journal of Law and Economics 32, 1-33.
Boubakri N., and C., Jean-Claude, 1998, The financial and operating performance of newly
privatized firms: Evidence from developing countries, Journal of Finance 53, 1081-1110.
Boycko, M., A. Shleifer and R. W. Vishny, 1996, A theory of privatization, Economic Journal
106, 309-319.
Byun, J., and M.S Rozeff, 2003, Long run performance after stock split: 1927 to 1996, Journal
of Finance 58, 1063-1085.
Brock, P.L., 2000, Emerging from Crisis: Bank privatization and recapitalization in Chile, in
Bank privatization; Conference proceedings of a Federal Reserve Bank of Dallas Policy
Research Workshop, ed by Harvey Rosenblum, 97-118.
Clarke, G.R.G., and R. Cull, 2002, Political and economic determinants of the likelihood of
privatizing Argentine public banks, Journal of Law and Economics, XLV, 165-197
Cornett, M. M, and H. Tehranian, 1990, An examination of the impact of the Garn-St Germain
Depository Institutions Act of 1982 on commercial banks and savings and loans, Journal
of Finance 45, 95-111.
Dewenter, K and P. Malatesta, 1997, Public offerings of state-owned and privately-owned
enterprises: an international comparison, Journal of Finance 52, 1659-1679.
31
Eckel, C., D. Eckel, and V. Singal, 1997, Privatization and efficiency: Industry effects of the
sale of British Airways, Journal of Financial Economics 43, 275-298.
Firth, M., 1996, Dividend changes, abnormal returns, and intra-industry firm valuations,
Journal of Financial and Quantitative Analysis, 31, 189-210.
Gunther, J.W and R. R. Moore, 2000, Financial statement and reality: Do troubled banks tell
all? Financial Industry Studies, Federal Reserve Bank of Dallas draft paper, Q3, 30 - 35.
Hertzel, M.G. 1991, The effect of stock repurchases on rival firms, Journal of Finance, 46,
707-15
La Porta R., F. Lopez-de-Silanes, and A. Shleifer, 2002, Government ownership of banks,
Journal of Finance, 57, 265– 301.
Laux, P., L. Stark and S. Y. Yoon, 1998, The Relative Importance of Competitive and
Contagion Intra-industry Information Transfers: An investigation of dividend
announcements, Financial Management, 27, 5 – 16.
Levis, M, 1993, The long-run performance of initial public offerings: The UK experience 19801988, Financial Management 22, 28-41.
Loughran, T and J. R., Ritter, and K. Rydqvist, 1994, The initial public offerings: International
insight, Pacific-Basin Finance Journal 2, 195-199.
Megginson, W. L., R. C. Nash, and M. van Randenborgh, 1994, The financial and operating
performance of newly privatized firms: An international empirical analysis, Journal of
Finance 49, 403-452.
Megginson, W. L., and J. M Netter, 1998, From state to market: A survey of empirical studies
on privatization, Journal of Economic Literature 39, 321-389.
Megginson, W. L., R.C. Nash, J. M Netter and A.L. Schwartz, 2000, The long run returns to
investors in share issue privatization, Financial Management 29, 67-77
Megginson, W. L
(2002) Sample firms privatized through public share offerings, 1961November
2002,
http://faculty-staff.ou.edu/M/William.L.Megginson-1,
University
of
Oklahoma
Menyah, K and K. Paudyal, 1996, Share issue privatization: The UK experience, in Mario
Levis (ed), Empirical Issues in Raising Equity Capital (Elsevier Science, Amsterdam).
Otchere, I and J. Chan, 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.
32
Persons, O. S., 1999, Using financial information to differentiate failed vs. surviving finance
companies in Thailand: An application for emerging economies, Multinational Finance
Journal 3,127-145.
Ritter, J. R., 1991, The long run performance of initial public offerings, Journal of Finance 46,
3-27.
Shleifer, A and R. Vishny, 1994, Politicians and Firms, Quarterly Journal of Economics 109,
995 – 1025.
Slovin, M., M. Sushka, and Y. Bendeck, 1991, The Intra-industry effects of going-private
transactions, Journal of Finance, September, 1537-1550.
Subrahmanyam A., and S. Titman (1999), The going-public decision and the development of
financial markets, Journal of Finance, 54, 1045-1082.
Szewczyck, S., 1992, The intra-industry transfer of information inferred from announcements
of corporate security offerings, Journal of Business, July, 1935-1945.
Thomson, J.B., 1991, Predicting bank failure in the 1980s, Economic Review-Federal Reserve
Bank of Cleveland 1st quarter, 9-20.
Verbrugge, J. A., W. L., Megginson, and W. Lee, 1999, The financial performance of
privatized banks: an empirical analysis, Working paper (University of Georgia).
33
Appendix A: Definition of ratios
Measures
Capital Adequacy
Ratios
Total Capital Adequacy
Calculation
=Tier 1 + Tier 2 Capital
Asset Quality
Provisions-to-Loans
Gross Impaired Assets-Loans
Net Impaired Assets to Loans
General provision / Total loans
Gross impaired assets /Total loans
Gross impaired assets less provisions/ Total Loans
Management Efficiency Cost-to-Income
Expense- to- Assets
Operating Expenses / Operating Income
Operating Expenses /Average Assets
Earnings Ability
Net Interest Margin
Return on Asset (ROA)
Return on Equity (ROE)
Net Interest income/Average interest earning assets
Net Profit Before Interest and Tax/Average Total Assets
Net Profit After Tax/Average Shareholders Equity
Employment
Growth in staff levels
Employee productivity
Over employment
% change in staff levels
Revenue (US$)/Number of employees
Total Assets/Number of employees
34
Table 1
Summary and descriptive statistics of the sample
The table presents frequency distribution and summary statistics for our sample of 18 privatized banks and 28 rival
banks from middle - and low-income countries. The full sample comprises 21 privatization and 68 rival banks
observations and it excludes privatized banks or rivals that did not have daily, monthly or income statement data.
Panel A presents the frequency distribution of the privatized banks and their rivals while Panel B shows the
summary statistics for the privatization transaction.
Panel A: Distribution of privatization *
Total observations
Privatized
Year
banks
Rivals
3
1989
1
1990
1991
1992
2
4
1993
2
3
1994
5
16
1995
4
18
1996
6
20
1997
1
1
Total
21
65
Rivals
Privatized banks
Country
Croatia
Egypt
Hungary
India
Jamaica
Kenya
Morocco
Philip pines
Poland
Total
No. of banks
1
1
1
1
1
2
4
1
6
18
Total
observations
1
1
1
2
1
2
4
2
7
21
No. of banks
4
3
1
4
2
2
5
4
3
28
Total
observations
4
3
1
6
2
4
20
7
18
65
Panel B: Summary statistics for the privatized banks
Issue size ($ US millions)
Percentage of capital offered
Percentage reserved for employees
Percentage reserved for foreign investors
Number of Shareholders created
Percentage owned by governments before privatization
Percentage owned by government after privatization
Initial return (%)
% of banks partially privatized
*
Mean
Median
156
27.69
7.28
29.88
459,739
75.97
49.20
90.12
88.1
66
25.75
6.25
29.30
20850
80
52.50
35
Extracted from Verbrugge et al (1999) and Megginson (2002).
35
Table 2
Rivals banks’ reaction to privatization announcement
This table presents the rival banks’ abnormal returns over different return intervals. Abnormal
returns are calculated using the market-adjusted model as the difference between the bank’s
returns and the return of the market index in the respective country. The abnormal returns are
cumulated over the –10 to +10 interval. The percentage negative is the ratio of firms with
negative abnormal returns to the total sample for the respective event windows.
Panel A: Abnormal returns for the whole sample
Mean Return (N=24)
Negative mean returns
Event period
[-10,+10]
[-5,+5]
[-2,+2]
[-1,+1]
[0]
[0, 1]
[0, 2]
[0, 3]
[0, 4]
[0, 5]
[1, 5]
[1, 10]
% Mean Return t-statistic
-1.93
-0.78
-1.94
-1.31
-0.32
-1.32
-2.24
-2.80
-2.73
-2.06
-1.74
-2.50
-1.14
-0.62
-1.58
-1.65
-0.85
-1.93*
-1.87*
-2.49**
-2.45**
-1.59
-1.31
-1.41
% Negative
71
67
54
54
54
54
54
63
58
67
58
54
z-statistic
2.04
1.63
0.41
0.41
0.41
0.41
0.41
1.22
0.82
1.63
0.82
0.41
Median Return
% Median returns z-statistic
-1.58
-0.46
-0.34
-0.42
-0.17
-0.03
-0.26
-1.00
-1.33
-1.72
-2.44
-1.25
1.24
1.10
1.13
0.90
0.75
1.07
1.13
1.81*
1.87*
1.73*
1.47
1.27
Panel B: Differential reaction of rival banks to privatization announcements
[-10,+10]
[-5,+5]
[-2,+2]
[-1,+1]
[0]
[0, 1]
[0, 2]
[0, 3]
[0, 4]
[0, 5]
[1, 5]
[1, 10]
***,**,*
% Mean Return
-3.04
-0.41
-2.46
-2.05
-0.34
-1.58
-2.26
-3.21
-2.96
-1.86
-1.52
-3.49
IPO
t-statistics
-0.96
-0.17
-1.24
-1.31
-0.53
-1.27
-1.17
-1.66
-1.54
-0.75
-0.59
-0.97
% Median Return % Mean Return
-5.76
-1.00
-0.28
-1.09
-1.65
-1.50
0.71
-0.69
0.00
-0.30
-0.01
-1.10
-0.44
-2.23
-1.38
-2.46
-1.39
-2.54
-2.42
-2.23
-2.62
-1.94
-2.65
-1.67
SEO
t-statistics
-0.58
-0.87
-0.95
-1.02
-0.66
-1.47
-1.42
-1.81*
-1.90*
-1.73*
-1.50
-1.23
% Median Return
-0.90
-1.15
0.01
-0.60
-0.19
-0.05
-0.07
-0.77
-1.27
-0.99
-1.21
0.64
Significant at 1%, 5% and 10% respectively
36
Table 3: Long run abnormal returns
This table contains mean and median long run returns for the privatized banks and rival banks. The return measures are marketadjusted cumulative abnormal returns and buy and hold returns (BAHR). The cumulative market-adjusted returns are from month 1 to
month 60 relative to the share issue month (month 0). The BAHR are the returns accruing to investors who bought the shares on the
first day of trading and held the stock for 12, 24, 36, 48 and 60 months. The figures in parentheses are t-statistics for mean returns or
z-statistics for the median returns or percentage positive. ***,**,* indicate significance at the 1%, 5% and 10% level respectively.
Panel A: Market-adjusted Returns
Privatized Banks
Event period
Mean Return % Positive
1-12
-9.61
36
(-0.82)
(-1.07)
1-24
-8.53
44
(-0.75)
(-0.47)
1-36
-15.01
53
(-0.94)
(0.23)
1-48
-16.31
43
(-1.05)
(-0.65)
1-60
-29.67
33
(-1.80)*
(-1.53)
Panel B: Buy and Hold Returns
Privatized Banks
Event period
Mean Return % Positive
1-12
1-24
1-36
1-48
1-60
-2.26
(-0.21)
-4.66
(-0.40)
-3.99
(-0.24)
-4.68
(-0.31)
-24.06
(-1.91)*
29
(-1.60)
47
(-0.26)
50
(0.00)
47
(-0.24)
25
(-2.00)*
Median Return Mean Return
-8.91
-1.59
(-0.94)
(-0.24)
-8.62
-9.53
(-0.70)
(-1.54)
4.01
-4.72
(0.87)
(-0.75)
-4.46
-3.31
(-0.87)
(-0.45)
-18.32
-10.34
(-1.60)
(-1.33)
Rival banks
Difference
% Positive Median Return Mean Return Median Return
46
-2.28
-8.02
-6.63
(-0.41)
(-0.47)
(-0.60)
(-0.94)
45
-3.51
1.00
-5.11
(-0.65)
(-1.33)
(0.08)
(-0.31)
55
8.33
-10.28
-4.32
(0.71)
(0.15)
(-0.60)
(-0.72)
50
0.42
-13.00
-4.88
(0.00)
(0.07)
(-0.76)
(-0.68)
46
-6.66
-19.32
-11.66
(-0.62)
(-1.45)
(-1.06)
(-1.11)
Median Return Mean Return
Rival banks
Difference
% Positive Median Return Mean Return Median Return
-14.38
(-0.69)
-7.17
(-0.20)
-8.00
(-0.28)
-7.20
(-0.19)
-31.26
(-1.73)*
-2.03
(-0.25)
-12.64
(-1.85)*
-6.49
(-0.95)
-0.89
(-0.10)
-4.51
(-0.42)
42
(-0.82)
41
(-1.15)
50
(0.00)
47
(-0.39)
39
(-1.66)*
-3.55
(-0.64)
-7.40
(-1.67)*
1.29
(0.44)
-1.58
(-0.63)
-10.51
(-2.17)**
-0.23
(-0.02)
7.98
(0.59)
2.50
(0.14)
-3.79
(-0.22)
-19.55
(-1.18)
-10.83
(-1.40)
0.23
(0.25)
-9.29
(-0.75)
-5.63
(-0.31)
-20.74
(-2.57)***
37
Table 4:
Operating performance measures
This table contains the median operating performance measures for the sample firms based on the CAMEL criteria. CAMEL stands
for Capital adequacy, Asset quality, Management efficiency, Earnings ability and Labor (employment levels and labor productivity).
Panel A presents the ratios of the privatized banks while Panel B exhibits the ratios of the rival banks. The z-statistics for the
Wilcoxon signed rank test of the difference in median ratios between the two samples are presented in Panel C. The median ratios and
z-statistics are reported where enough observations exist to compute the z-statistic.
Capital
Adequacy
Year
Tier 1+Tier 2
-3
-2
-1
0
1
2
3
4
5
12.15
14.46
13.60
14.80
14.20
14.13
13.85
12.50
12.65
-3
-2
-1
0
1
2
3
4
5
17.07
18.65
15.80
12.00
11.10
11.39
11.70
12.40
Asset Quality
Gross Impaired
Provisions-to- Assets-toLoans
Loans
Management Efficiency
Net Impaired
Assets -toLoans
Cost-toIncome
Panel A: Median Ratios of Privatized Banks
2.85
21.54
16.78
5.83
15.47
12.84
3.26
13.74
10.43
3.06
17.02
13.66
1.80
20.72
13.66
1.74
13.92
11.47
0.87
9.95
9.10
1.09
8.85
8.52
0.77
10.47
9.84
Panel B: Median Ratios of Rival Banks
1.04
0.48
0.74
1.43
6.73
2.05
1.37
6.59
5.19
1.32
6.73
6.12
0.94
7.33
6.24
1.20
8.58
7.67
1.20
9.79
8.72
Earnings Ability
Expense-to- Net Interest
Assets
Margin
Labor
ROE
ROA
Growth in Employee Asset -tostaff levels Productivity Employee
83.51
76.83
78.98
74.14
75.68
76.56
80.16
82.51
82.81
16.36
17.48
14.00
15.08
13.59
12.92
13.30
12.96
12.94
1.11
7.22
3.88
8.25
7.09
5.72
4.78
3.97
4.87
22.38
19.85
13.80
16.32
23.67
21.17
20.27
17.13
14.29
1.84
2.82
1.61
3.43
3.47
3.13
2.37
2.54
2.23
3.20
2.80
1.31
0.98
1.31
2.16
1.29
2.87
3.65
0.11
0.11
0.09
0.08
0.09
0.09
0.09
0.09
0.08
0.58
0.52
0.34
0.34
0.45
0.45
0.45
0.54
0.51
79.13
79.71
79.20
79.20
77.87
77.11
78.23
79.11
81.15
11.53
9.31
16.51
12.38
10.27
10.59
10.28
9.56
10.01
0.13
0.14
0.60
6.18
4.91
4.91
4.74
4.55
4.19
29.62
25.62
18.81
17.13
16.86
17.13
16.19
15.97
12.18
2.68
2.46
2.77
2.49
2.48
2.46
2.44
1.98
1.81
0.77
2.89
2.66
2.66
2.36
1.06
1.33
0.06
0.05
0.06
0.08
0.09
0.09
0.09
0.09
0.09
0.45
0.24
0.36
0.77
1.10
0.95
0.75
0.68
0.68
38
Table 4 continued
Panel C: z-statistics for difference in median (Privatized Banks – Rivals)
-3
1.60
0.37
1.60
-2
-1.84*
2.02**
-1.69*
1.46
-1
-2.39**
1.96**
-0.36
-0.17
0
-1.01
2.05**
2.37**
2.81***
-2.33**
1.77*
1
0.66
1.89*
3.52***
3.60***
-1.17
0.71
2
2.74***
2.75***
2.99***
3.04***
-0.15
1.95*
3
4.04***
-2.78***
2.36**
2.71***
1.34
4.40***
4
2.31**
-1.49
0.45
0.46
2.58***
5.65***
5
1.46
-3.74***
0.24
0.98
1.22
5.35***
***,**,* significant at 1%, 5% and 10% respectively (2-tailed test)
1.60
1.46
0.68
2.64***
2.79***
1.56
0.70
-2.56***
4.70***
-1.60
-1.46
-1.69*
-1.24
2.79***
1.73*
2.63***
2.00**
1.15
-1.60
1.10
-2.37**
2.20**
2.76***
3.22***
-0.63
3.67***
3.38***
0.45
-0.95
-1.44
-2.07**
-2.17**
1.72*
2.40**
1.60
1.84*
1.96**
0.26
0.60
0.86
0.04
0.26
-2.04**
39
1.07
1.84*
-1.16
-2.06**
-2.91***
-4.30***
-5.17***
-3.91***
-4.34***
Table 5:
Difference in mean test
This table shows the mean pre- and post-privatization ratios and the associated t-statistics for the sample based on the CAMEL
criteria. CAMEL stands for Capital adequacy, Asset quality, Management efficiency, Earnings ability and Labor (employment levels
and labor productivity). The mean pre-privatization ratios are calculated over the year –3 to year –1 period and the mean postprivatization ratios are calculated over year +1 to year +5, relative to year 0 (the year of issue). Lack of data precluded the calculation
of mean pre privatization gross- and net-impaired assets-to-loans ratios for the rival banks.
Privatized Banks Mean performance
Ratio
Capital Adequacy
Provisions-to-Loans
Gross Impaired Assets to Loans
Net Impaired Assets to Loans
Cost- to- Income
Expense-to-Asset
Net interest margin
ROA
ROE
Growth in Staff
Employee Productivity
Assets to Employees
Post
Pre
Post-Pre
t-statistic
Pre-privatization
privatized
Rivals
Post-privatization
Diff
t-statistic
privatized
Rivals
Diff
t-statistic
13.58
18.45
-4.88
- 0.89
18.45
17.53
0.93
0.17
13.58
11.76
1.81
3.23***
2.30
3.91
-1.61
- 2.14**
3.91
1.94
1.97
2.32**
2.30
1.35
0.95
2.06*
14.76
15.18
-0.42
- 0.07
15.18
-
-
-
14.76
9.33
5.43
1.77*
11.84
10.13
1.71
0.30
10.13
-
-
-
11.84
8.17
3.68
1.41
78.67
77.58
1.10
0.28
77.58
81.83
-4.25
- 0.96
78.67
80.69
- 2.02
- 0.67
12.25
14.65
-2.40
- 1.22
14.65
14.66
- 0.02
- 0.01
12.25
10.36
1.89
2.03**
5.12
5.08
0.04
0.03
5.08
3.96
1.12
0.53
5.12
4.70
0.42
0.68
2.78
2.76
0.02
0.02
2.76
2.84
- 0.07
- 0.10
2.78
2.06
0.72
0.93
14.43
17.51
-3.08
- 0.45
17.51
20.82
- 3.31
- 0.59
14.43
12.62
1.81
0.38
1.51
5.33
-3.82
- 0.65
5.33
1.54
3.78
0.68
1.51
3.34
-1.83
- 0.72
0.10
0.11
-0.01
- 0.24
0.11
0.07
0.04
0.80
0.10
0.10
0.00
- 0.27
0.16
0.74
0.50
0.48
0.02
0.10
0.66
0.93
- 0.26
- 2.11**
0.66
0.50
***,**,* significant at 1%, 5% and 10% respectively
40
Table 6:
Determinants of long run stock market performance
The table presents OLS regression results of the long run returns (CAR1,60) on firm
specific factors, as well as characteristics of the privatization transactions. Size is the
logarithmic of the mean post-privatization total revenue; LowD is an indicator variable
equal to one if the firm is in a low-income country and zero of it is in a middle-income
country; %Gov is the percentage of government ownership of the privatized bank after
privatization; FullD is the full privatization dummy and it takes on a value of 1 if the
transaction is full privatization and 0 if it is partial privatization; Prod is the mean post
privatization employee productivity variable; CTI is the mean post privatization mean
cost-to-income ratio; SEOD is an indicator variable equal to one if the transaction is a
subsequent equity offering and zero if it is an IPO, and LLP is the mean post-privatization
loan loss provisions ratio. The t-statistics are based on White heteroskedasticityconsistent standard errors. ***,**,* indicate that the coefficient is significantly different
from zero at 1%, 5% and 10% respectively.
Privatized Banks
Independent Variables
Intercept
Size
LowD
LLP
FullD
%Gov
SEOD
CTI
Prod
2
Adjusted R
N
F
Rival Banks
Coefficient
0.87
0.0001
t-statistics
1.13
3.12**
Coefficient
1.82
-6E-05
t-statistics
2.33**
-0.44
0.70
-0.01
-1.55
-0.05
-2.00
0.02
7.89
1.82
-0.11
-8.12***
-6.71***
-5.39***
1.54
2.10*
-0.91
0.25
-0.84
-0.004
0.07
-0.02
-2.27
-3.64***
1.65
-4.15***
-1.21
0.38
-2.56**
-1.38
0.88
13
12.42**
0.28
42
3.03**
41
Figure 1: Cumulative market adjusted returns
0.2
CAR
59
57
55
53
51
49
47
45
43
41
39
37
35
33
31
29
27
25
23
21
19
17
15
13
9
11
7
5
3
0
1
0.1
-0.1
Month
-0.2
-0.3
-0.4
-0.5
Privatized
Rivals
42