Differences between Politically

Connected and Nonconnected Firms:

A Cross-Country Analysis
Mara Faccio∗

Evidence from firms in 47 countries shows that companies with political connections have higher
leverage and higher market shares, but they underperform compared to nonconnected compa-
nies on an accounting basis. Differences between connected and unconnected firms are more
pronounced when political links are stronger. Differences also vary depending on the level of
corruption and the degree of economic development in individual countries.

Despite the fact that corruption has a negative aggregate economic effect on a country’s
investment and growth, a growing body of literature has pointed out that political connections may
be beneficial to specific firms. Academic studies reporting evidence on how connections provide
sources of value have identified only a few differences between connected and nonconnected
firms, such as preferential access to credit (Johnson and Mitton, 2003; Chiu and Joh, 2004;
Dinç, 2004; Cull and Xu, 2005; Khwaja and Mian, 2005), government contracts (Goldman,
Rocholl, and So, 2008), regulatory protection (Kroszner and Stratmann, 1998), and government
aid for financially troubled firms (Faccio, Masulis, and McConnell, 2006). Additionally, most
of these studies look at individual countries and highly dissimilar types of connections, making
cross-country comparison virtually impossible.1
By contrast, the purpose of this paper is to analyze how connected firms differ from noncon-
nected firms across a large number of countries. I have two main questions: First, are differences
between connected and nonconnected firms common across countries, or are they specific to a
few countries and/or types of connections? Second, are these differences larger in countries with
high corruption and/or in less developed countries?
To address these questions, I use a new database built in Faccio (2006) that includes several
thousand firms in 47 countries. A company is defined as connected with a politician if “at least
one of its large shareholders (anyone controlling at least 10 percent of voting shares) or one of its

Some of the results in this paper were previously circulated with the title “Politically-Connected Firms: Can They Squeeze
the State?” I thank two anonymous referees, Massimo Belcredi, Bernardo Bortolotti, Lorenzo Caprio, Tom Cosimano,
Serdar Dinç, Simeon Djankov, Ray Fisman, Mariassunta Giannetti, Richard Green, Tom Gresik, E. Han Kim, Meziane
Lasfer, Ron Masulis, Paolo Mauro, John McConnell, Todd Mitton, Randall Morck, David Parsley, Susan Rose-Ackerman,
Paul Schultz, Andrei Shleifer, and David Stolin for their helpful comments, and Suzanne Bellezza, Pam Losefsky and
Patricia Peat for editorial help. I also benefited from comments from workshop participants at Indiana University,
Stockholm School of Economics, Università Cattolica (Milan), Università di Lugano, University of Michigan, University
of Notre Dame, Syracuse University, Vanderbilt University, the IMF, the AFA meeting, and at the METU International
Conference in Economics. I also wish to thank Larry Lang for sharing data on group affiliation in Asia.

Mara Faccio is the Hanna Chair in Entrepreneurship and Associate Professor of Finance at the Krannert School of
∗

Management at Purdue University in West Lafayette, IN.

1
The lone exception is Faccio, Masulis, and McConnell (2006).
Financial Management • Autumn 2010 • pages 905 - 927
906 Financial Management ! Autumn 2010
top officers (CEO, president, vice-president, chairman, or secretary) is a member of parliament,
a minister, or is closely related to a top politician or party” (Faccio, 2006, p. 369). As recognized
in Faccio (2006), relying on publicly available data sources yields an incomplete picture of
connections. However, it seems reasonable to believe that connections identified through public
sources are more likely to represent durable ties (as opposed to those related to ephemeral
campaign contributions).
I find that connected firms have higher leverage, pay lower taxes, and have stronger market
power; however, they have poorer accounting performance than nonconnected firms.2 They
differ more dramatically from their peers when their political links are stronger. Differences are
greater, for example, when companies are connected through owners rather than through directors.
Similarly, they are greater when the connection is with a minister, rather than with a member of
the parliament. I additionally find that the financial characteristics of connected firms differ more
from those of their nonconnected peers in countries characterized by high levels of corruption.
Also, the differences between connected and nonconnected firms vary across countries depending
on economic development, for example, the level of per capita GDP. Results are robust to the
exclusion of outliers and individual countries and industries, and to a number of other tests.
There are two possible interpretations of these results: 1) it could be that political connections
led to these differences or 2) it could be that these types of firms are more likely to establish
political connections. To the extent that the characteristics studied reflect benefits of connections
across countries, these results indicate that the distortions in the allocation of public resources
often claimed by practitioners are common in both emerging and developed countries. However,
the magnitude of this phenomenon is much larger in more corrupt systems. This last result
is consistent with Faccio (2006) who, for a large sample of connections across developed and
emerging markets, documents that firm value increases when an entrepreneur is elected to a top
political position, especially in highly corrupt countries. Overall, the evidence suggests caution
in making inferences about the benefits of political ties because the magnitude of the benefits
depends on the specific country of analysis.
This work complements a growing body of literature studying the net effect of political con-
nections on the value of firms. This evidence has largely shown that, on average, the benefits of
having political connections exceed the costs. For example, Fisman’s (2001) study of connections
to Indonesian President Suharto shows that rumors of Suharto’s worsening health significantly
(and negatively) affected the prices of companies related to the president. Studies by Roberts
(1990) and Goldman, Rocholl, and So (2009) provide evidence that connections through board
members, on average, add to the value of US firms.3 Ramalho (2003) and Ferguson and Voth
(2008) provide similar evidence for firms related to Brazilian President Collor de Mello and
German firms affiliated with the Nazi party (during its rise to power), respectively. In related
work, Aggarwal, Meschke, and Wang (2007) and Gulen, Cooper, and Ovtchinnikov (2010) look
at the impact of political contributions on long-term stock returns.
The rest of the paper is organized as follows. In Section I, I define political connections and
in Section II present evidence on the characteristics of connected firms. Section III discusses the
cross-country evidence. Section IV provides a number of robustness tests. Section V concludes
the paper.

2
This may either proxy preferential access to the credit market or for the possibility that these firms have lower costs of
default or lower probability of default due to their connections.
3
Fisman et al. (2006), however, report that the value of companies tightly connected to Vice President Cheney is not
affected by major health events involving the vice president, or major political news.
Faccio ! Differences between Politically Connected and Nonconnected Firms 907

I. Definition of Political Connections

Data on political ties come from Faccio (2006), who analyzes political connections across 47
countries. In particular, a company is defined as politically connected if “at least one of its large
shareholders (anyone controlling at least 10 percent of voting shares) or one of its top officers
(CEO, president, vice-president, chairman, or secretary) is a member of parliament, a minister,
or is closely related to a top politician or party” (Faccio, 2006, p. 369).
Connections with government ministers include cases in which the politician himself is a
director or a large shareholder, as well as cases where a politician’s close relative holds such a
position.
Close relationships include: 1) companies whose top executives or large shareholders have been
described in the press as having a friendship with a head of state, government minister, or member
of parliament; 2) connections with officials who had served as heads of state or prime ministers
in the past; 3) companies whose former top executives or large shareholders entered politics; 4)
connections with foreign politicians; and 5) other connections identified in prior studies (Gomez
and Jomo, 1997; Johnson and Mitton, 2003).
Finally, I include cases in which a member of parliament serves as a company’s CEO, president,
vice president, or secretary, or controls at least 10% of shareholder votes. Because of data
limitations, I record parliamentary connections only when the members of parliament themselves
are shareholders or top directors, but do not consider cases when parliamentary positions are held
by relatives.
To establish the presence of connections, I identify names of members of parliament or gov-
ernment using the Chiefs of State directory (CIA, 2001) and the official website of each country’s
government and parliament. I take the names of top company directors from Worldscope, Extel,
company websites, and LexisNexis. I identify major shareholders from Claessens, Djankov, and
Lang (2000), Faccio and Lang (2002), the websites of the stock exchanges or their supervisory
authorities, Worldscope, and Extel. The Economist, Forbes, and Fortune are the sources of infor-
mation for well-known cases of friendships. Additionally, I use studies by Agrawal and Knoeber
(2001), Backman (1999), Gomez and Jomo (1997), Johnson and Mitton (2003), and Fisman
(2001) to identify close relationships of the types listed above.
The data set built in Faccio (2006) identifies 541 companies with political connections in 47
countries. To be included in my sample, financial information for these companies needs to be
available in Worldscope for 1997. Because of financial data availability issues, this requirement
reduces the sample to 458 connected firms and 15,733 nonconnected peers. Overall, there are
514 connections, as some companies have multiple ties. Of these, 307 connections (60%) involve
top directors, while 207 cases (40%) involve large shareholders. In the majority of cases (304,
or 59%), connections are with members of parliament, followed by close relationships (132
cases, or 26%), and connections with government ministers or heads of state (78 cases, or 15%).
Table I shows the country distribution of connections.

II. The Differences between Politically Connected and

Nonconnected Firms
This section studies the differences between politically connected and nonconnected firms, in
terms of: 1) leverage, 2) taxation, 3) market power, 4) productivity, 5) accounting performance,
and 6) market-to-book ratio. Although given the limitations of the data, I cannot infer causal-
ity from my results, in principle, I attempt to focus on characteristics that are more likely to
908 Financial Management ! Autumn 2010
Table I. Country Distribution of Firms with Political Connections

No. of firms with available data is the number of firms with financial data available in Worldscope. No.
of connected firms is the number of firms whose controlling shareholder or top director sits on a national
parliament, government, is king/president of the country, or is closely related to a top politician/political
party.

Country No. of Firms No. of Country No. of Firms No. of

with Available Connected with Available Connected
Data Firms Data Firms
Argentina 34 0 New Zealand 47 0
Australia 257 2 Norway 113 0
Austria 87 1 Peru 24 0
Belgium 104 5 Philippines 100 5
Brazil 127 0 Poland 32 0
Canada 438 6 Portugal 60 3
Chile 71 2 Russian Fed. 11 4
Czech Rep. 58 0 Singapore 215 16
Denmark 172 7 South Africa 188 0
Finland 91 2 South Korea 271 7
France 519 19 Spain 138 1
Germany 507 10 Sweden 172 3
Greece 90 1 Switzerland 180 4
Hong Kong 381 7 Thailand 204 32
Hungary 26 1 Turkey 78 1
India 257 8 United Kingdom 1,417 119
Indonesia 116 27 United States 6,007 13
Ireland 52 2 Venezuela 17 0
Israel 47 2 Colombia 32 0
Italy 178 21 Luxembourg 23 1
Japan 2,322 30 Sri Lanka 18 0
Malaysia 418 81 Taiwan 237 7
Mexico 68 7 Zimbabwe 8 0
Netherlands 179 1
All countries 16,191 458

reflect benefits of connections such as preferential access to credit, tax discounts, and market
protection/monopolies.
Before proceeding with the analysis, it is important to point out that at least three factors work
against finding any statistical differences between politically connected and nonconnected firms.
First, to the extent that high leverage or low taxation reflects a benefit, this benefit may accrue
mostly to unlisted firms connected with politicians. Since financial data are not widely available
for unlisted firms, I cannot test this hypothesis. Second, benefits may be granted industrywide
rather than to specific firms. This is often true in the case of barriers to entry and tax relief.
Stigler (1971) discusses several such cases in the United States. Third, since many connected
firms may operate as monopolies or quasi-monopolies, their industry-adjusted financial ratios
will be exactly the same as those of their “peers.”
Table II provides preliminary univariate statistics for the sample of connected firms and their
nonconnected peers. (All variables are defined later in this section). It shows that the leverage of
connected firms is significantly higher than that of their nonconnected peers (28.14% vs. 24.19%).
Faccio ! Differences between Politically Connected and Nonconnected Firms 909
Table II. Descriptive Statistics

Connected companies are those whose controlling shareholder or top director sits on a national parliament,
government, is king/president of the country, or is closely related to a top politician/political party. Leverage
is defined as long-term debt (excluding the current portion of long-term debt, pensions, deferred taxes, and
minority interest) over total capital × 100. Total capital represents the total investment in the company: the
sum of common equity, preferred stock, minority interest, long-term debt, nonequity reserves, and deferred
tax liability in untaxed reserves. Tax is income taxes over pretax income × 100. Market share is firm’s
market capitalization over the total market capitalization of all firms in the same country and two-digit SIC
industry (%). Productivity is the natural log of total-factor productivity estimated assuming a Cobb-Douglas
production function. Return on assets (ROA) is the ratio of a company’s net income prior to financing costs
to total assets (×100). Market-to-book is the ratio of market value of (ordinary and preferred) equity plus
the book value of debt, divided by the sum of book value of equity plus book value of debt. Mkcap is the
company’s market capitalization, defined as market price as of year end × common shares outstanding (in
millions of US$). State represents the voting stake held by the central and local government. It is calculated
by identifying the weakest link in each control chain linking the corporation to the controlling shareholder,
then summing the percentage control rights across these links. Privatized is a dummy that equals 1 if the
company is a privatized firm, and 0 otherwise. Dually listed is a dummy that equals 1 if the company is listed
on at least two stock markets, and 0 otherwise. The t-test tests for the equality of means between connected
and nonconnected companies. The Kruskal-Wallis test tests for the equality of the medians between these
two groups. A test on the equality of proportions is used for dummy variables.

Connected Nonconnected t-Test Kruskal-Wallis Test of

(p-Value) Test Proportion
Mean Median Mean Median (p-Value) (p-Value)
Leverage 28.14 22.67 24.19 18.62 0.00 0.00
Tax 29.67 30.00 32.70 34.43 0.00 0.00
Market share 18.04 5.26 9.48 0.61 0.00 0.00
Productivity 0.74 0.73 0.80 0.79 0.05 0.00
ROA 2.91 3.41 3.30 3.94 0.67 0.06
Market-to-book 2.09 1.29 2.77 1.46 0.66 0.00
Mkcap ($M) 3,634.65 309.41 1,265.04 166.85 0.00 0.00
State (%) 1.79 0.00 0.88 0.00 0.00 0.04
Privatized (%) 4.37 1.33 0.00
Dually listed (%) 21.40 20.78 0.75

Connected firms also enjoy significantly lower tax rates (29.67% vs. 32.7%) and have larger
market shares (18.04% vs. 9.48%). The productivity of connected firms is significantly lower
than the productivity of nonconnected peers. However, on average, the accounting performance
and market-to-book ratios are not different from a statistical standpoint. Connected firms are
substantially larger (three times larger, on average, in terms of market value of equity), they have
more government ownership, and they are more likely to be former state-owned enterprises.

A. Leverage
Leverage is a proxy for access to debt financing. Leverage is defined as the ratio of long-term
debt (excluding the current portion of long-term debt, pensions, deferred taxes, and minority
interest) to total capital × 100. Total capital represents the total investment in the company. It is
the sum of common equity, preferred stock, minority interest, long-term debt, nonequity reserves,
and deferred tax liability in untaxed reserves.
910 Financial Management ! Autumn 2010
To investigate the relationship between connectedness and leverage, I initially pool all connec-
tions together (Table III , Panel A); next, I categorize them as either: 1) connections through the
owner and connections through a director (Panel B) or 2) connections with the king, president,
or minister, or connections with members of parliament and/or close relationships (Panel C).
Unless otherwise specified, all regressions control for whether the company is recently pri-
vatized, state controlled, or dually listed, as well as size (market capitalization), country, and
industry, defined according to Campbell (1996).4 I identify dual listings (e.g., whether a com-
pany is listed on at least two stock markets) from Worldscope. I obtain lists of privatized firms
from SDC Platinum, Bortolotti, Fantini, and Siniscalco (2001), Dewenter and Malatesta (1997),
and Megginson, Nash, and Van Randenborgh (1994), and update these with data kindly provided
by Professor Megginson.5 I use Extel, Worldscope, Claessens, Djankov, and Lang (2000), Faccio
and Lang (2002), and the 2000 “Fortune 500 Global List” to identify government ownership. The
“typical baseline” (OLS) regression being performed is

Leveragei = α + β1 Connected i + β2 Privatized i + β3 State-controlled i

C
! K
!
+ β4 Dually listed i + β5 Sizei + δc Countryc + γk Industryk + u i . (1)
1 1

Each Panel in Table III refers to a different set of regressions, using different measures of
connections. An alternative approach would be to look at changes in leverage ratios (as well as
taxation and market share) before and after the initial date the connection was made. However, a
precise event date can be identified for only a small proportion of firms.6
Connected firms have significantly higher leverage than nonconnected ones. Furthermore,
leverage is marginally higher when connections are stronger. For example, the excess leverage
is marginally higher for firms connected through their owner, than for firms connected through
a director. Leverage is the highest in cases of close relationships (coeff. = 7.209); next highest
for firms connected with the king, the president, or a minister (coeff. = 3.597); and lowest for
connections with a member of parliament (coeff. = 1.035). Results are robust to the exclusion of
financial companies (see Table III, Panel B).
While connected firms are more levered than nonconnected ones, they do not necessarily enjoy
a benefit in the form of reduced costs of debt financing. For the whole sample, the average interest
rate on debt (interest paid/total debt) is only marginally lower for connected firms (a difference of
−0.07%) and far from significance (this result is not formally reported in a table). For companies
connected with a minister, however, the average interest rate on debt is lower by 1.14 percentage
points (p-value = 0.05), supporting the view that connections with more influential politicians
are worth more.

4
Industries are defined as follows: petroleum (SIC 13, 29), consumer durables (SIC 25, 30, 36-37, 50, 55, 57), basic
industry (SIC 10, 12, 14, 24, 26, 28, 33), food and tobacco (SIC 1-2, 9, 20-21, 54), construction (SIC 15-17, 32, 52),
capital goods (SIC 34-35, 38), transportation (SIC 40-42, 44-45, 47), utilities (SIC 46, 48-49), textiles and trade (SIC
22-23, 31, 51, 53, 56, 59), services (SIC 72-73, 75-76, 80, 82, 87, 89), leisure (SIC 27, 58, 70, 78-79), and financial
companies (SIC 60-69).
5
Choi, Lee, and Megginson (2010) provide recent evidence on the long-run performance of privatization IPOs.
6
For such a subsample, Boubakri, Cosset, and Saffar (2009) document that leverage increases significantly after the
establishment of a new political connection.
Faccio ! Differences between Politically Connected and Nonconnected Firms 911
Table III. Characteristics of Connected Firms

The table reports the coefficients for ordinary least squares regressions where the variables are defined
as follows. The dependent variables are in boldface. Total capital represents the total investment in the
company: the sum of common equity, preferred stock, minority interest, long-term debt, nonequity reserves,
and deferred tax liability in untaxed reserves. Tax is income taxes over pretax income × 100. Market share
is firm’s market capitalization over the total market capitalization of all firms in the same country and
two-digit SIC industry (%). Productivity is the natural log of total-factor productivity estimated assuming a
Cobb-Douglas production function. Return on assets (ROA) is the ratio of a company’s net income prior to
financing costs to total assets (×100). Market-to-book is the ratio of market value of (ordinary and preferred)
equity plus the book value of debt, divided by the sum of book value of equity plus book value of debt.
Connected is a dummy that equals 1 if the company’s controlling shareholder or top director sits on a national
parliament, government, is king/president of the country, or is closely related to a top politician/political
party, and 0 otherwise. Connected through the owner is a dummy that equals 1 if the company’s controlling
shareholder sits in a national parliament, holds office in the government, is the head of state, or is closely
related to a top politician/political party, and 0 otherwise. Connected through a director is a dummy that
equals 1 if a company’s top director sits in a national parliament, holds office in the government, is the
head of state, or is closely related to a top politician/political party, and 0 otherwise. Connected with king,
president, or minister is a dummy that equals 1 if a controlling shareholder or top director of the company
holds a government office, or is king/president of the country, and 0 otherwise. Connected with MP is a
dummy that equals 1 if a controlling shareholder or top director of the company sits in a national parliament,
and 0 otherwise. Close relationships is a dummy that equals 1 if a controlling shareholder or top director of
a company is closely related to at least one top politician, and 0 otherwise. Close relationships include: 1)
companies whose top executives or large shareholders have been described in the press as having a friendship
with a head of state, government minister, or member of parliament; 2) connections with officials who had
served as heads of state or prime ministers in the past; 3) companies whose former top executives or large
shareholders entered politics; 4) connections with foreign politicians; and 5) other connections identified in
prior studies (Gomez and Jomo, 1997; Johnson and Mitton, 2003). Industry is defined according to Campbell
(1996). Standard errors (reported in parentheses below the coefficients) are computed using Huber/White
correction for heteroskedasticity (see White, 1980).

Panel A. General Results

Leverage Tax Market Productivity ROA Market-to-

Share Book
Connected 3.014∗∗∗ −0.715 2.198∗∗ −0.021 −2.410∗∗∗ −0.482∗
(1.146) (0.859) (1.104) (0.021) (0.517) (0.286)
Privatized 3.719∗∗ −1.735 12.806∗∗∗ −0.009 −3.847∗∗∗ −1.589
(1.705) (1.109) (2.117) (0.023) (0.633) (0.987)
State controlled 1.097 0.457 8.228∗∗ −0.027 −1.666 −2.605
(3.922) (2.272) (3.739) (0.080) (1.154) (1.720)
Dually listed 3.798∗∗∗ −0.981∗∗ −0.382 −0.021∗∗ 0.221 0.423
(0.527) (0.405) (0.388) (0.008) (0.425) (1.068)
Size 0.711∗∗∗ 0.674∗∗∗ 2.917∗∗∗ 0.023∗∗∗ 1.581∗∗∗ 0.366∗∗
(0.114) (0.086) (0.094) (0.002) (0.111) (0.183)
Intercept 4.176 32.624∗∗∗ −16.189∗∗∗ 0.060 −16.576∗∗∗ −3.994
(3.847) (8.883) (2.268) (0.072) (1.780) (2.476)
Country dummies Y Y Y Y Y Y
Industry dummies Y Y Y Y Y Y
R2 adj. 11.51% 27.71% 37.40% 8.81% 6.34% 1.86%
N. Obs. 16,138 12,304 16,147 11,551 15,687 16,143

(Continued)
912 Financial Management ! Autumn 2010
Table III. Characteristics of Connected Firms (Continued)

Leverage Tax Market Productivity ROA Market-to-

Share Book

Panel B. Director versus Shareholder Connection

Connected 3.475∗∗ −2.457∗ 2.825 0.051 −2.994∗∗∗ −0.111

through the (1.742) (1.419) (1.914) (0.049) (0.888) (0.337)
owner
Connected 1.948 0.272 1.875 −0.051∗∗ −1.485∗∗ −0.659∗∗
through a (1.438) (1.005) (1.274) (0.022) (0.587) (0.337)
director
Privatized 3.771∗∗ −1.799 12.838∗∗∗ −0.008 −3.894∗∗∗ −1.578
(1.705) (1.111) (2.117) (0.023) (0.633) (0.987)
State controlled 1.134 0.361 8.268∗∗ −0.024 −1.708 −2.591
(3.921) (2.274) (3.742) (0.080) (1.153) (1.720)
Dually listed 3.794∗∗∗ −0.974∗∗ −0.386 −0.021∗∗∗ 0.225 0.422
(0.527) (0.405) (0.387) (0.008) (0.425) (1.068)
Size 0.715∗∗∗ 0.674∗∗∗ 2.916∗∗∗ 0.024∗∗∗ 1.578∗∗∗ 0.367∗∗
(0.114) (0.086) (0.094) (0.002) (0.111) (0.183)
Intercept 4.154 32.737∗∗∗ −16.167∗∗∗ 0.060 −16.494∗∗∗ −3.973
(3.845) (8.893) (2.268) (0.072) (1.887) (2.472)
Country dummies Y Y Y Y Y Y
Industry dummies Y Y Y Y Y Y
R2 adj. 11.51% 27.73% 37.41% 8.83% 6.34% 1.86%
N. Obs. 16,138 12,304 16,147 11,551 15,687 16,143

Panel C. Connections with Members of Parliament versus Connections with Ministers

Connected with 3.597 0.955 −3.247 0.171∗ −3.044∗ 0.362

king, president, (3.402) (2.376) (2.930) (0.091) (1.671) (0.606)
or minister
Connected with 1.035 −1.021 1.882 −0.028 −1.469∗∗ −0.753∗∗
MP (1.441) (0.997) (1.252) (0.022) (0.575) (0.361)
Close 7.209∗∗∗ −0.734 5.766∗∗ −0.030 −4.281∗∗∗ −0.304
relationships (2.088) (2.014) (2.512) (0.071) (1.074) (0.395)
Privatized 3.796∗∗ −1.730 12.835∗∗∗ −0.008 −3.886∗∗∗ −1.580
(1.704) (1.109) (2.114) (0.023) (0.631) (0.986)
State controlled 1.162 0.470 8.220∗∗ −0.026 −1.704 −2.594
(3.916) (2.272) (3.736) (0.080) (1.153) (1.720)
Dually listed 3.788∗∗∗ −0.985∗∗ −0.385 −0.021∗∗ 0.225 0.422
(0.527) (0.405) (0.388) (0.008) (0.425) (1.068)
Size 0.708∗∗∗ 0.675∗∗∗ 2.912∗∗∗ 0.023∗∗∗ 1.582∗∗∗ 0.367∗∗
(0.114) (0.087) (0.094) (0.002) (0.111) (0.183)
Intercept 4.203 32.613∗∗∗ −16.145∗∗∗ 0.060 −15.854∗∗∗ −4.124∗
(3.842) (8.926) (2.267) (0.072) (1.817) (2.473)
Country dummies Y Y Y Y Y Y
Industry dummies Y Y Y Y Y Y
R2 adj. 11.54% 27.71% 37.44% 8.82% 6.36% 1.86%
N. Obs. 16,138 12,304 16,147 11,551 15,687 16,143
∗∗∗
Significant at the 0.01 level.
∗∗
Significant at the 0.05 level.
∗
Significant at the 0.10 level.
Faccio ! Differences between Politically Connected and Nonconnected Firms 913
B. Taxation
The variable tax is defined as the ratio of Income Taxes/Pretax Income × 100. In the calculation
of the tax variable, I exclude companies with negative earnings and companies that display a tax
rate above 100%.7
The results indicate that connected firms enjoy lower taxation than their nonconnected peers.
The difference between the tax rate of connected versus unconnected firms, however, is not
statistically significant. Results are significant only for the one subgroup of connected firms that
displays relatively stronger connections: companies connected through their owners. Surprisingly,
connections with the king, president, or a minister are associated with insignificantly higher
taxation. It could be that connections to the king are obtained because a firm is willing to pay
higher taxes, as paying higher taxes directly to the king gets the firm connections that lead to
other private benefits.
One problem with analyzing taxation is that tax breaks may be granted industrywide, rather
than for one connected firm, leading to insignificant differences across firms. To assess this
possibility, I rerun all tax regressions in Tables III through VI, eliminating industry dummies
(results not reported for space reasons). Results are essentially unchanged after the exclusion of
industry dummies. A second limitation is that, while my tax variable captures tax breaks in the
form of special deductions, it does not reflect tax savings that occur when companies are allowed
to underreport income. Third, if corporate tax rates are graduated in some countries, then tax rates
could be low for connected firms not because they’re getting favors from the government, but
because their profitability is lower. Typically, firms in countries in which tax rates are graduated
tend to depend on the level of pretax profits. Thus, a way to address this concern is to add a control
for the firm’s pretax income in the tax regressions. (These results are not formally reported in
a table). Interestingly, pretax income turns out to be negatively correlated with tax rates. More
importantly, the results on the connection variables maintain the same sign as those in Table III,
with only “connections to the owners” being associated with significantly lower tax rates.

C. Market Share
Market share may come either from a real monopolistic or quasi-monopolistic position, from
some advantage in obtaining concessions or licenses, or from sizable government contracts.
Alternatively, large firms may find connections to be more valuable, and, therefore, be more
likely to establish them. In other words, the stakes may be higher for firms that have monopoly
power; therefore, these firms may have more incentive to establish connections.
I measure Market share as the firm’s market capitalization as a proportion of the total market
capitalization of all firms in the same country and two-digit SIC industry (percent). I use market
cap instead of sales because the sample includes financial companies. However, results are similar
if I use sales instead of market capitalization and exclude financial companies.
The evidence on the market share variable is also strong. Connected firms enjoy a significantly
higher market share than nonconnected peers. In particular, market share is especially higher
when the connection is through close relationships (coeff. = 5.766). These results confirm that

7
Tax rates above 100% might simply reflect data errors. At the same time, it might indeed be the case that the estimated
tax rates are above 100%. One possible reason for the estimated tax rates to exceed 100% relates to the difference between
reported income and taxable income, which is the case in many countries. Consider a country with a marginal corporate
income tax rate of 10%. If a firm with a reported profit of $100 is reporting costs for $1,000 that are not immediately tax
deductible, its reported income ($100) will be lower than its taxable income ($1,100). This firm will pay taxes for 110
(10% of $1,100), corresponding to a tax rate of 110%.
914 Financial Management ! Autumn 2010
stronger connections are associated with, at least at the margin, sharper differences between
connected and nonconnected firms.

D. Productivity
Are connected firms more or less productive than nonconnected companies? To estimate
productivity, I assume a standard Cobb-Douglas production function
β γ
Yi = Pi K iα L i Mi , (2)

where Yi is firm i’s output, measured by its revenues (in US$); K i are inputs of capital, measured
by total assets (in US$); L i are inputs of labor, measured by the number of employees; and Mi
are inputs of materials, measured by the cost of goods sold (in US$). Pi is firm i’s total-factor
productivity. α, β, and γ are the output elasticities of capital, labor, and inputs, respectively. To
estimate Pi , I first take natural logs of Equation (2),

yi = pi + αki + βli + γ m i + εi , (3)

where pi , ki, li , and mi are the natural logs of total-factor productivity and inputs of capital, labor,
and materials, respectively, and εi is an error term. I then use ordinary least squares (OLS) to
estimate Equation (3), and solve it for pi . Thus, the estimated Productivity, p̂i , is calculated as

p̂i = yi − α̂ki − β̂li − γ̂ m i . (4)

In general, connected companies appear to be as productive as their nonconnected counterpar-

ties. However, productivity is significantly lower when the connection is through a director. On
the other hand, productivity is higher when the connection is with the king, the president, or a
minister. However, these results are overall mixed, because in some instances stronger connections
are associated with lower, and in other instances they are associated with higher productivity.

E. Accounting Performance and Market Valuation

Here, there are two measures of company performance/value of interest:

– Return on assets (ROA) is the ratio of a company’s net income prior to financing costs to
total assets (×100).
– Market-to-book is computed as the ratio of market value of (ordinary and preferred) equity
plus the book value of debt, divided by the sum of book value of equity plus book value of
debt.

One might expect connected firms to report better performance because of benefits obtained
from connections. Furthermore, better performing firms may become connected to maintain their
power and performance. Poorly performing companies may become connected for two reasons.
First, they may seek connections as a way to obtain relief from some of their problems. Second,
firms owned or managed by politicians may be poor performers because their managers lack the
skills needed to run a successful company. In both cases, connected firms may underperform
even though connections may be value enhancing. Connected firms might also have to devote
substantial resources to their rent-seeking activities, which may well eliminate any advantage
Faccio ! Differences between Politically Connected and Nonconnected Firms 915
from the connections they have (see Fisman, 2001; Johnson and Mitton, 2003). De Soto (1989)
argues that, in Peru, bribes to curry favor replace the taxes that companies do not pay. Shleifer
and Vishny (1994) note that relationships between politicians and firms have a price. While
politicians will be willing to provide subsidies to firms run by independent managers, they will
want firms to pay them back by pursuing social policy goals.8
Results for my measure of company performance indicate that connected firms are poor
performers. The ROA of connected firms is lower by 2.41 percentage points (p-value < 0.01),
and their market-to-book ratio is lower by 0.48 (difference only marginally significant). It is not
clear which type of connection is associated with lower profitability. All subgroups of connected
firms exhibit significantly lower ROA than their nonconnected peers. Companies connected
through the weakest relationships (e.g., those connected through a director, and those connected
with a member of parliament) have the lowest market valuation.
Given the existing evidence that connections add value, my interpretation is that the poor
accounting performance of connected firms likely reflects ex ante underperformance.9 In other
words, the performance of these companies is low even though political connections, on average,
add value to the sample firms. This interpretation is consistent with the evidence in Boubakri,
Cosset, and Saffar (2009) who, for a subsample of politically connected firms whose precise
connection date could be identified, show that the performance of connected firms increases
significantly subsequent to the establishment of a connection.

III. Cross-Country Evidence

A. Country-Level Results
Analysis of country-level results is important because the benefits of political connections may
vary by country. In the British system, a member of Parliament has no unusual power because
voting occurs by party, while representatives in the United States vote independently on many
issues. Because there are just a handful of connected firms in many countries, I focus on the
top countries in terms of: 1) number of connected firms, 2) proportion of politically connected
listed firms, or 3) connected firms as a proportion of the market capitalization (omitting Ireland
because it has only two connected firms).
The results are in Table IV. For all except Italy, connected firms display higher leverage.
Leverage is significantly higher for connected firms in Malaysia, Russia, Thailand, and the
United Kingdom. In Italy, leverage is not only lower, but significantly so. In all seven countries,
connected firms are subject to marginally lower rates of taxation, although the results always lack
statistical significance at conventional levels.
For five of the seven countries, connected firms have higher market share. This relationship
is statistically significant and economically large for the Russian sample. The productivity of

8
To test for the possible level of overemployment by connected firms, I compute the ratio of the number of employees
over total assets. I regress this ratio against the connections proxies used above, as well as various control factors. In
no case is there a significant difference in the level of employment by connected versus nonconnected firms (results
not reported to save space). The difference is generally very small in economic terms (the coefficient of the connection
dummy is −0.001), far from significant, and generally has the wrong sign, that is, connected firms employ fewer people.
The data do not allow examination of other interesting sources of costs across countries, for example, the remuneration of
politicians who sit on boards of directors or payments made to politicians, whether legally (i.e., campaign contributions)
or illegally.
9
See, for example, Aggarwal, Meschke, and Wang (2007), Faccio (2006), Ferguson and Voth (2008), Goldman, Rocholl,
and So (2009), Gulen, Cooper, and Ovtchinnikov (2010), Ramalho (2003), and Roberts (1990).
916 Financial Management ! Autumn 2010
Table IV. Country-Level Regressions

The table reports the coefficients for ordinary least squares (OLS) regressions where the variables are
defined as follows. The dependent variables are in boldface. All (OLS) regressions control for whether the
firm has recently been privatized, and whether it is state controlled, dually listed, operates in the financial
industry (SIC between 6000 and 6999), as well as for firm size (ln{mkcap}). All regressions include an
intercept. Coefficients for these control variables are not reported to save space. Leverage is defined as
long-term debt (excluding the current portion of long-term debt, pensions, deferred taxes, and minority
interest) over total capital × 100. Total capital represents the total investment in the company: the sum
of common equity, preferred stock, minority interest, long-term debt, nonequity reserves, and deferred
tax liability in untaxed reserves. Tax is income taxes over pretax income × 100. Market share is firm’s
market capitalization over the total market capitalization of all firms in the same country and two-digit SIC
industry (%). Productivity is the natural log of total-factor productivity estimated assuming a Cobb-Douglas
production function. Return on assets (ROA) is the ratio of a company’s net income prior to financing costs
to total assets (×100). Market-to-book is the ratio of market value of (ordinary and preferred) equity plus
the book value of debt, divided by the sum of book value of equity plus book value of debt. Connected is a
dummy that equals 1 if the company’s controlling shareholder or top director sits on a national parliament,
government, is king/president of the country, or is closely related to a top politician/political party, and 0
otherwise. Standard errors (reported in parentheses below the coefficients) are computed using Huber/White
correction for heteroskedasticity (see White, 1980). Horizontal lines separate different regressions, and each
column refers to a different regression.

Leverage Tax Market Share Productivity ROA Market-to-Book

Indonesia
Connected 2.273 −0.516 −3.433 0.034 −6.117∗ −0.178
(5.868) (5.854) (6.713) (0.099) (3.511) (0.142)
R2 adj. 13.09% 6.52% 10.61% 34.91% 14.17% 31.37%
N. Obs. 116 66 116 34 115 116
Italy
Connected −10.726∗∗ −3.927 0.350 0.028 −1.222∗ −0.525
(5.089) (3.599) (5.411) (0.160) (0.707) (0.433)
2
R adj. 25.85% 4.49% 23.15% 12.26% 31.22% 18.96%
N. Obs. 178 149 178 128 176 178
Japan
Connected 0.304 −4.079 2.054 −0.031 0.444 0.241
(4.310) (2.629) (3.203) (0.033) (0.689) (0.149)
R2 adj. 2.80% 1.60% 15.84% 3.66% 9.16% 9.99%
N. Obs. 2,322 1,786 2,322 2,156 2,322 2,322
Malaysia
Connected 10.708∗∗∗ −0.792 0.967 0.029 −0.956 0.189
(2.546) (2.581) (3.109) (0.135) (1.575) (0.768)
R2 adj. 6.54% 8.01% 17.30% 19.30% 2.08% 5.39%
N. Obs. 418 300 418 41 398 418
Russia
Connected 9.233∗∗∗ −73.266 67.662∗∗ −0.445∗∗ −6.129∗∗ 0.612
(2.457) (44.209) (20.245) (0.102) (1.348) (1.185)
R2 adj. 73.73% 64.91% 55.55% 51.45% 89.68% 54.33%
N. Obs. 11 8 11 8 6 11

(Continued)
Faccio ! Differences between Politically Connected and Nonconnected Firms 917
Table IV. Country-Level Regressions (Continued)

Leverage Tax Market Share Productivity ROA Market-to-Book

Thailand
Connected 17.266∗∗ −3.020 −6.081 −0.226 −7.018∗∗ −0.059
(6.785) (4.724) (4.452) (0.374) (2.846) (0.243)
R2 adj. 10.46% 4.24% 32.58% 24.78% 5.05% 5.61%
N. Obs. 204 119 204 24 204 204
United Kingdom
Connected 3.566∗ −1.478 2.819 −0.018 −1.451 −0.704∗∗
(2.094) (1.161) (1.735) (0.030) (1.030) (0.341)
R2 adj. 5.58% 4.91% 23.21% 0.65% 2.75% 4.57%
N. Obs. 1,417 1,200 1,417 1,155 1,405 1,417
∗∗∗
Significant at the 0.01 level.
∗∗
Significant at the 0.05 level.
∗
Significant at the 0.10 level.

connected firms is lower than the productivity of nonconnected companies in four out of seven
countries; however, this difference is statistically significant only in Russia. Connected firms
exhibit lower ROA in all countries but Japan; the difference is significant in Indonesia, Italy, Rus-
sia, and Thailand. Firms in the United Kingdom exhibit significantly lower market-to-book ratio.
Connected firms (insignificantly) outperform nonconnected peers (in market-to-book terms) in
Japan, Malaysia, and Russia.

B. Where Are Differences Greatest?

To test whether connections are more important in countries with higher levels of corruption,
or in less developed ones, I analyze interaction dummies between connections and corruption, as
well as interactions between connections and per capita GDP.
As a proxy for corruption, I use the average of four indexes used elsewhere as measures of
perceived corruption: the Business International index; the International Country Risk index;
the Kaufmann, Kraay, and Zoido-Lobatón index (http://www.worldbank.org/wbi/governance/
datasets.html#dataset); and the Transparency International index (www.transparency.org).10 Cor-
ruption indexes are rescaled from 0 to 10, so that lower scores correspond to lower levels of
corruption. The Business International’s (Economist Intelligence Unit) index assesses the “degree
to which business transactions involve corruption and questionable payments.” This assessment
is compiled based upon questionnaires filled in by Business International’s network of corre-
spondents and analysts based in the countries covered and reflect their perception of corruption.
The International Country Risk index assessing the corruption in government indicates whether
“high government officials are likely to demand special payments” and whether “illegal payments
are generally expected throughout lower levels of government” in the form of “bribes connected
with import and export licenses, exchange controls, tax assessment, policy protection, or loans.”
The Kaufmann, Kraay, and Zoido-Lobatón Corruption index is defined as the exercise of public
power for private gains and measures various aspects, ranging from the frequency of “additional
payments to get things done” to the effects of corruption on the business environment. “The
indicator reflects the statistical compilation of perceptions of the quality of governance of a large

10
The same results obtain if the individual indexes are used.
918 Financial Management ! Autumn 2010
number of survey respondents in industrial and developing countries, as well as non-governmental
organizations, commercial risk rating agencies, and think-tanks during 1997 and 1998.” Finally,
Transparency International measures the “degree to which corruption is perceived to exist among
public officials and politicians. It is a composite index, drawing on 14 different polls and surveys
from seven independent institutions, carried out among business people and country analysts,
including surveys of residents, both local and expatriate.” Corruption represents “the abuse of
public office for private gain.” Ln{GDP (per capita)} is the natural log of gross domestic product
(GDP) (in US$) on a purchasing power parity basis divided by population, computed for 1999
(Source: World Bank, http://sima-ext.worldbank.org/query/).
It is particularly important to control for per capita GDP to avoid spurious findings, as corrup-
tion is correlated with many other country-level variables. For example, the correlation between
corruption and per capita GDP is −0.77. At the same time, the inclusion of highly correlated vari-
ables may induce multicollinearity problems, so that the estimated coefficients for the variables
of interest may be biased. To address multicollinearity issues, corruption and per capita GDP
are orthogonalized using a modified Gram-Schmidt procedure. As a result of this procedure, the
corruption and per capita GDP variables employed in the analysis are demeaned.
Table V shows that the earlier documented differences between connected and nonconnected
firms are particularly sharp in highly corrupt systems. The results are more mixed with respect
to per capita GDP, as, for some variables, lower per capita GDP widens the differences, while
for other variables, a lower per capita GDP shrinks the differences between connected and
nonconnected firms.
In the leverage regression, the coefficient of the interaction between the general connection
variable and corruption is 2.405 (significant at the 1% level). The tax rates of connected firms are
also lower in more corrupt countries, although insignificantly so. The market share of connected
firms increases insignificantly with corruption.
Interestingly, the productivity of connected firms decreases significantly with the level of
corruption in the country. Similarly, the accounting performance of connected firms significantly
deteriorates with the level of corruption in the country, while their market-to-book ratio does
not change in a significant way. In countries with higher corruption, this may suggest that while
connected firms are relatively inefficient, connections provide the benefits that allow them to
more than compensate for any lack of management skills.
Table V also confirms that the type and strength of connections matters. For market share
a connection through the owner, rather than through a director, has a greater marginal effect
when corruption increases. (Curiously, the leverage regression shows lower leverage for compa-
nies connected through the owner). Firms connected through close relationships exhibit greater
differences in their leverage, taxation, and market share as corruption increases.

IV. Robustness Tests

A. Various Empirical Tests
I perform a number of robustness tests on the results. First, I split the sample between countries
with corruption above the sample median (Table VI, Panel A) and countries with corruption
equal to or below the median (Panel B). The excess leverage of connected firms is significant
in highly corrupt countries; however, market share is significant only in countries with relatively
low corruption. As before, taxation and productivity are never significantly different between
connected and nonconnected firms. For both samples, the ROA of connected firms is significantly
Faccio ! Differences between Politically Connected and Nonconnected Firms 919
Table V. Differences across Countries

The table reports the coefficients for ordinary least squares regressions where the variables are defined as
follows. The dependent variables are in boldface. Each column refers to a different regression. Leverage is
defined as long-term debt (excluding the current portion of long-term debt, pensions, deferred taxes, and
minority interest) over total capital × 100. Total capital represents the total investment in the company:
the sum of common equity, preferred stock, minority interest, long-term debt, nonequity reserves, and
deferred tax liability in untaxed reserves. Tax is income taxes over pretax income × 100. Market share
is firm’s market capitalization over the total market capitalization of all firms in the same country and
two-digit SIC industry (%). Productivity is the natural log of total-factor productivity estimated assuming
a Cobb-Douglas production function. ROA is the ratio of a company’s net income prior to financing costs
to total assets (×100). Market-to-book is the ratio of market value of (ordinary and preferred) equity plus
the book value of debt, divided by the sum of book value of equity plus book value of debt. Connected is a
dummy that equals 1 if the company’s controlling shareholder or top director sits on a national parliament,
government, is king/president of the country, or is closely related to a top politician/political party, and 0
otherwise. Connected through the owner is a dummy that equals 1 if the company’s controlling shareholder
sits in a national parliament, holds office in the government, is the head of state, or is closely related to a
top politician/political party, and 0 otherwise. Connected through a director is a dummy that equals 1 if a
company’s top director sits in a national parliament, holds office in the government, is the head of state,
or is closely related to a top politician/political party, and 0 otherwise. Connected with king, president,
or minister is a dummy that equals 1 if a controlling shareholder or top director of the company holds a
government office, or is king/president of the country, and 0 otherwise. Connected with MP is a dummy
that equals 1 if a controlling shareholder or top director of the company sits in a national parliament, and
0 otherwise. Close relationships is a dummy that equals 1 if a controlling shareholder or top director of
a company is closely related to at least one top politician, and 0 otherwise. Close relationships include:
1) companies whose top executives or large shareholders have been described in the press as having
a friendship with a head of state, government minister, or member of parliament; 2) connections with
officials who had served as heads of state or prime ministers in the past; 3) companies whose former
top executives or large shareholders entered politics; 4) connections with foreign politicians; and 5) other
connections identified in prior studies (e.g., Gomez and Jomo, 1997; Johnson and Mitton, 2003). Ave.
corruption is the average of four indexes used elsewhere as measures of perceived corruption: the Business
International index, the International Country Risk index, the Kaufmann, Kraay, and Zoido-Lobatón index
(http://www.worldbank.org/wbi/governance/datasets.html#dataset), and Transparency International index
(www.transparency.org). Corruption indexes are rescaled from 0 to 10, so that lower scores correspond
to lower levels of corruption. Ln{GDP (per capita)} is the natural log of gross domestic product (GDP)
(in US$) on a purchasing power parity basis divided by population; computed for 1999 (Source: World
Bank, http://sima-ext.worldbank.org/query/). Because of multicollinearity issues, the corruption and per
capita GDP employed have been orthogonalized using a modified Gram-Schmidt procedure. Standard
errors (reported in parentheses below the coefficients) are computed using Huber/White correction for
heteroskedasticity (see White, 1980).

Panel A. General Results

Leverage Tax Market Productivity ROA Market-to-

Share Book
Connected −0.192 −0.177 3.671∗∗∗ −0.013 −1.749∗∗∗ −1.033∗∗
(1.439) (1.057) (1.295) (0.025) (0.499) (0.436)
Connected × ave. 2.405∗∗∗ −0.849 0.077 −0.030∗ −1.614∗∗∗ −0.545
corruption (0.846) (0.675) (0.824) (0.018) (0.422) (0.529)
Connected × −2.109∗ −1.323 6.201∗∗∗ 0.018 1.555∗∗∗ 0.063
Ln{GDP (1.108) (0.900) (1.043) (0.025) (0.439) (0.214)
(per capita)}

(Continued)
920 Financial Management ! Autumn 2010
Table V. Differences across Countries (Continued)

Leverage Tax Market Productivity ROA Market-to-

Share Book

Panel A. General Results (Continued)

Ave. corruption 1.311∗∗∗ −0.296∗ 3.916∗∗∗ −0.057∗∗∗ 0.777∗∗∗ 0.533

(0.210) (0.162) (0.256) (0.005) (0.141) (0.585)
Ln{GDP (per capita)} 1.681∗∗∗ 4.966∗∗∗ −6.202∗∗∗ 0.054∗∗∗ −2.196∗∗∗ −0.364
(0.193) (0.143) (0.215) (0.004) (0.135) (0.229)
Additional Controls: Recently Privatized, State Controlled, Dually Listed, Size and Industry Dummies
R2 adj. 6.42% 13.07% 20.57% 5.81% 4.29% 0.15%
N. Obs. 16,138 12,304 16,147 11,551 15,687 16,143

Panel B. Director versus Shareholder Connection

Connected through −7.584∗∗∗ −1.409 7.505∗∗∗ 0.011 −1.637 −0.831

the owner (2.211) (1.844) (2.843) (0.058) (1.050) (0.567)
Connected through 1.998 1.005 0.815 −0.039 −1.294∗∗ −0.872∗∗
a director (1.763) (1.367) (1.407) (0.024) (0.560) (0.346)
Connected through 4.368∗∗∗ −0.810 0.167 −0.014 −1.719∗∗ −0.584
the owner × ave. (1.278) (1.096) (1.332) (0.026) (0.727) (0.510)
corruption
Connected through 0.824 −0.648 −0.270 −0.051∗ 0.133 −0.468
a director × ave. (1.437) (1.182) (1.352) (0.029) (0.628) (0.466)
corruption
Connected through −4.574∗∗∗ −2.417∗∗ 8.621∗∗∗ 0.059 2.426∗∗∗ 0.025
the owner × (1.402) (1.136) (1.768) (0.052) (0.524) (0.289)
Ln{GDP (per
capita)}
Connected through −0.163 0.112 3.734∗∗∗ 0.001 −0.040 0.222
a director × (1.457) (1.342) (1.307) (0.026) (0.542) (0.223)
Ln{GDP (per
capita)}
Ave. corruption 1.329∗∗∗ −0.302∗ 3.897∗∗∗ −0.057∗∗∗ 0.772∗∗∗ 0.531
(0.209) (0.162) (0.256) (0.005) (0.141) (0.583)
Ln{GDP (per capita)} 1.681∗∗∗ 4.958∗∗∗ −6.192∗∗∗ 0.054∗∗∗ −2.193∗∗∗ −0.365
(0.193) (0.143) (0.214) (0.004) (0.135) (0.229)
Additional Controls: Recently Privatized, State Controlled, Dually Listed, Size and Industry Dummies
R2 adj. 6.47% 13.08% 20.61% 5.85% 4.31% 0.15%
N. Obs. 16,138 12,304 16,147 11,551 15,687 16,143

(Continued)

worse than the ROA of nonconnected firms. However, this poorer accounting performance does
not result in below-market valuation when the connected firm operates in a country with higher
corruption.
Panel C of Table VI excludes Malaysia and Indonesia, as Fisman (2001) and Johnson and
Mitton (2003) have already documented substantial effects associated with connections in those
countries. The basic results continue to hold after excluding these two countries. Exclusion of the
United Kingdom and the United States (with most firms) does not change the results, either.
Faccio ! Differences between Politically Connected and Nonconnected Firms 921
Table V. Differences across Countries (Continued)

Leverage Tax Market Productivity ROA Market-to-

Share Book

Panel C. Connections with Members of Parliament versus Connections with Ministers

Connected with king, −0.991 2.520 3.322 0.128 −4.422∗∗ −0.138

president, or (7.075) (3.715) (3.938) (0.126) (2.009) (0.687)
minister
Connected with MP −0.256 −0.624 2.058 −0.010 −0.855∗ −1.092∗∗∗
(1.623) (1.207) (1.406) (0.027) (0.518) (0.411)
Close relationships −0.187 0.092 10.114∗∗∗ −0.106 −5.663∗∗∗ −2.143
(3.387) (2.857) (3.850) (0.081) (1.350) (1.459)
Connected with king, 1.967 −1.220 −2.980 0.000 −0.539 −0.991
president, or (2.057) (1.575) (1.969) (0.045) (1.155) (0.815)
minister × ave.
corruption
Connected with MP× 0.932 −1.596 −1.347 −0.009 0.045 −0.480
ave. corruption (1.560) (1.362) (1.076) (0.015) (0.680) (0.509)
Close relationships × 2.062∗∗∗ −0.735∗∗∗ 1.347∗∗∗ 0.004 −0.025 0.192
ave. corruption (0.165) (0.147) (0.212) (0.003) (0.115) (0.170)
Connected with king, 0.397 0.844 3.610∗ 0.096 1.340 −0.018
president, or (4.657) (1.786) (2.131) (0.084) (1.141) (0.604)
minister ×
Ln{GDP (per
capita)}
Connected with −2.362∗ −0.720 6.484∗∗∗ −0.005 0.297 0.059
MP × Ln{GDP (1.330) (1.340) (1.112) (0.021) (0.551) (0.227)
(per capita)}
Close relationships × −5.536∗∗ −1.512 8.755∗∗∗ 0.054 2.079∗∗ −0.145
Ln{GDP (per (2.371) (2.250) (2.860) (0.073) (1.003) (0.448)
capita)}
Ave. corruption −1.536∗∗∗ 0.652∗∗∗ 2.028∗∗∗ −0.063∗∗∗ 0.793∗∗∗ 0.256
(0.278) (0.227) (0.373) (0.006) (0.232) (0.401)
Ln{GDP (per capita)} 3.501∗∗∗ 4.323∗∗∗ −5.013∗∗∗ 0.058∗∗∗ −2.220∗∗∗ −0.196
(0.237) (0.189) (0.286) (0.005) (0.174) (0.187)
Additional Controls: Recently Privatized, State Controlled, Dually Listed, Size and Industry Dummies
R2 adj. 7.46% 13.33% 21.13% 5.85% 4.32% 0.16%
N. Obs. 16,138 12,304 16,147 11,551 15,687 16,143
∗∗∗
Significant at the 0.01 level.
∗∗
Significant at the 0.05 level.
∗
Significant at the 0.10 level.

Panel D excludes financial firms (SIC 6000-6999). Nonfinancial connected companies con-
tinue to exhibit significantly higher leverage than their peers and significantly greater market
shares. Once again, they significantly underperform on an accounting basis.
Panel E excludes government-controlled and privatized firms, whose objectives may be other
than to maximize shareholder value and which may exhibit abnormal ratios. Results are qualita-
tively unchanged after this exclusion. Results do not appear to be driven by outliers (Panel F). In
fact, key results are robust to truncating the data at the 5th and 95th percentiles of observations
922 Financial Management ! Autumn 2010
Table VI. Robustness Tests

The table reports the coefficients for ordinary least squares regressions where the variables are defined
as follows. The dependent variables are in boldface. Each column refers to a different regression. With
the exception of Panel E, all regressions control for whether the firm is politically connected, has recently
been privatized, is state controlled, dually listed, as well as for firm size (ln{mkcap}), country and industry
dummies. Regressions in Panel E, however, do not control for whether the firm has recently been privatized
or is state controlled. Industry is defined according to Campbell (1996). Coefficients for the control variables
are not reported to save space. Leverage is defined as long-term debt (excluding the current portion of long-
term debt, pensions, deferred taxes, and minority interest) over total capital × 100. Total capital represents
the total investment in the company: the sum of common equity, preferred stock, minority interest, long-term
debt, nonequity reserves, and deferred tax liability in untaxed reserves. Tax is income taxes over pretax
income × 100. Market share is firm’s market capitalization over the total market capitalization of all firms in
the same country and two-digit SIC industry (%). Productivity is the natural log of total-factor productivity
estimated assuming a Cobb-Douglas production function. Return on assets (ROA) is the ratio of a company’s
net income prior to financing costs to total assets (×100). Market-to-book is the ratio of market value of
(ordinary and preferred) equity plus the book value of debt, divided by the sum of book value of equity
plus book value of debt. Connected is a dummy that equals 1 if the company’s controlling shareholder
or top director sits on a national parliament, government, is king/president of the country, or is closely
related to a top politician/political party, and 0 otherwise. Standard errors (reported in parentheses below
the coefficients) are computed using Huber/White correction for heteroskedasticity (see White, 1980).

Leverage Tax Market Share Productivity ROA Market-to-Book

Panel A. Countries with Corruption above the Sample Median
∗∗∗
Connected 4.530 −2.313 0.280 −0.027 −2.023∗∗∗ −0.666
(1.715) (1.468) (1.615) (0.037) (0.729) (0.630)
R2 adj. 15.01% 39.42% 39.26% 24.06% 23.63% 2.72%
N. Obs. 5,241 3,927 5,244 3,459 5,141 5,242

Panel B. Countries with Corruption Equal to or Below the Sample Median

Connected 1.563 0.661 3.801∗∗∗ −0.010 −2.438∗∗∗ −0.467∗∗

(1.499) (0.974) (1.469) (0.025) (0.746) (0.200)
R2 adj. 10.36% 14.89% 35.62% 6.12% 4.92% 0.30%
N. Obs. 10,897 8,377 10,903 8,092 10,546 10,901

Panel C. All Countries Except Malaysia and Indonesia

Connected 2.345∗ −0.138 2.349∗∗ −0.021 −2.095∗∗∗ −0.638∗∗

(1.304) (0.901) (1.195) (0.022) (0.532) (0.304)
R2 adj. 11.50% 28.78% 38.22% 8.31% 6.45% 1.86%
N. Obs. 15,604 11,938 15,613 11,476 15,174 15,609

Panel D. All Industries Except Financial Services (SIC: 6000-6999)

Connected 2.498∗∗ −1.192 3.028∗∗ −0.018 −2.356∗∗∗ −0.552

(1.238) (1.017) (1.431) (0.021) (0.648) (0.433)
R2 adj. 13.25% 28.66% 40.05% 9.33% 8.13% 3.25%
N. Obs. 12,887 9,548 12,893 10,996 12,538 12,891

(Continued)
Faccio ! Differences between Politically Connected and Nonconnected Firms 923
Table VI. Robustness Tests (Continued)

Leverage Tax Market Share Productivity ROA Market-to-Book

Panel E. All Companies Except Government-Controlled and Privatized Firms

Connected 3.061∗∗ −0.397 2.087∗ −0.025 −2.466∗∗∗ −0.567

(1.202) (0.914) (1.107) (0.023) (0.561) (0.345)
R2 adj. 11.00% 28.25% 36.22% 8.81% 6.26% 1.97%
N. Obs. 15,395 11,692 15,404 11,102 14,952 15,400

Panel F. Dependent Variable Truncated at the 5th and 95th Percentiles

Connected 1.655∗ −0.325 1.993∗∗∗ −0.024∗ −1.165∗∗∗ −0.106∗∗

(0.974) (0.647) (0.628) (0.013) (0.304) (0.048)
R2 adj. 10.88% 30.76% 37.11% 19.21% 15.38% 26.25%
N. Obs. 15,331 11,075 14,573 10,399 14,116 14,529

Panel G. All Countries with at Least One Connected Firm

Connected 2.930∗∗ −0.726 2.195∗∗ −0.020 −2.433∗∗∗ −0.473

(1.146) (0.858) (1.105) (0.021) (0.518) (0.289)
R2 adj. 11.14% 27.32% 34.37% 8.12% 6.17% 1.88%
N. Obs. 15,458 11,759 15,465 11,162 15,019 15,463
∗∗∗
Significant at the 0.01 level.
∗∗
Significant at the 0.05 level.
∗
Significant at the 0.10 level.

for the dependent variable. Finally, the results do not appear to be driven by the inclusion of a
few countries with no politically connected firms (Panel G).

B. Direction of Causality: Anecdotal Evidence

As, because of data limitations, I employ OLS regressions throughout the analysis, it is not
possible to infer any causality from the results reported. However, anecdotal evidence is more
supportive of the idea that high leverage, low taxes, and larger market share reflect benefits (e.g.,
are the consequence) of connections, rather than being the cause. I report some of these interesting
cases in this section.
Anecdotal evidence suggests that connected firms enjoy easier access to debt financing from
state-controlled banks, even though they are not worth this extra credit. For example, in 1982,
a company owned by Daim Zainuddin (former Malay Deputy Prime Minister and close friend
of Prime Minister Mahathir), Baktimu Sdn Bhd, acquired a 33% stake in Sime UEP, for RM 75
million cash. “Part of the loan for the acquisition, amounting to RM 40m, was obtained from
the Singapore branch of the Union Bank of Switzerland; the loan was approved by the Union
Bank only after the government-owned Bank Bimiputra issued a guarantee on Bakrimu’s behalf
as security for the credit” (Asian Wall Street Journal, August 24, 1984; Gomez and Jomo, 1997,
pp. 54-55). In the aftermath of the Asian financial crisis, when the Indonesian Bank Restructuring
Agency published the names of Indonesia’s main corporate borrowers, Suharto’s children figured
on the top of the list: “Second on the list, with 3.5 trillion rupiah in loans, is Timor Putra Nasional,
the auto firm controlled by Tommy Suharto. Number five, with 2.9 trillion rupiah in debt, is a
924 Financial Management ! Autumn 2010
petrochemical company owned by the timber tycoon Prajogo Pangestu and Suharto’s second
son, Bambang Trihatmodjo.”11 In 1986, François Pinault, the controlling shareholder of Pinault
SA (France) obtained a 250 million FF grant from the government (US$40 million). By 1997,
Crédit Lyonnais’ credits and stakes in Pinault had reached a value of 12 billion FF (US$2.14
billion) (Calvi and Meurice, 1999; Gay and Monnot, 1999). Similarly, Italian Prime Minister
Silvio Berlusconi was accused of financing his television empire through the “large helping hand
[of] public-sector banks, which provided bigger loans than Fininvest’s creditworthiness seemed
to merit” (The Economist, 2001).
Similarly, evidence suggests that connections lead to preferential tax treatment. In 1996, Russian
President Boris Yeltsin signed a decree giving tax breaks and other aid potentially worth more
than US$1 billion to Norilsk Nickel, one of the country’s richest and most influential industrial
giants. Norilsk was controlled by Uneximbank, whose president, Vladimir Potanin, was shortly
thereafter appointed deputy prime minister (The Moscow Times, 1996). Similarly, when in 1996
Pinault SA obtained the cash contribution from the French government, it was also given a tax
exemption of 250 million FF (Gay and Monnot, 1999).
More generally, in Indonesia, a condition for IMF lending in the aftermath of the Asian financial
crisis was that Suharto sign an agreement compelling companies controlled by his family to give
up lucrative government concessions, monopolies, licenses, government contracts, and tax breaks
that protected them from competition in their domestic market.12 Such benefits were found to be
huge.
Anecdotes on market power are legion. As described in Backman (1999, pp. 266-268), “money
from the [Suharto] family’s start-up capital came from having themselves granted import monop-
olies. One of the earliest such monopolies was an exclusive license for the import of raw materials
for plastic, granted in 1984.” Similarly, Malay crony capitalists are rent-seeking “private sector
businessmen who benefit enormously from close relations” with government leaders by obtaining
“not only protection from foreign competition, but also concessions, licenses, monopoly rights,
and government subsidies” (Yoshihara, 1988, pp. 3-4). Relationships became so widespread that
by 1995 almost 20% of the Malay ruling party’s division chairmen were millionaire businessmen
(Gomez and Jomo, 1997, p. 26). In the Philippines, connected firms could easily obtain licenses
by paying a 10% fee (Hutchcroft, 1998, p. 73).

V. Conclusion

Several findings on the relationship between politics and business are revealed in this examina-
tion of connected corporations across several countries. Connected companies differ sharply from
those not connected. On average, leverage is higher in connected corporations. Connected firms
also enjoy marginally lower taxation, and they display much greater market power. Connected
firms display lower ROA and market valuation than their peers. These results are generally con-
sistent across countries. This evidence is consistent with, and complements the results in previous
studies of preferential access to credit by politically connected firms. For example, Khwaja and
Mian (2005) show that connected firms in Pakistan enjoy greater access to debt financing, and

11
Arnold, Wayne, “Indonesia’s repo man: Eko Budianto has ordered corporate cronies from the Suharto regime to pay
back the billions they owe Indonesian banks or he’ll seize their assets, even if it means enlisting the army to help him,”
The New York Times, July 31, 1999.
12
Rosenthal, A. M., “What Suharto Knew,” The New York Times, March 10, 1998. Pura, Raphael, “Rising Resentment:
Scrutiny of Suharto Wealth Is Intensifying in Jakarta,” The Asian Wall Street Journal, June 1, 1998.
Faccio ! Differences between Politically Connected and Nonconnected Firms 925
although they exhibit significantly higher default rates, they pay no higher interest rates than
their nonconnected peers.13 To the best of my knowledge, however, this is the first study to
systematically document tax discounts or strong market positions among connected owners.
I also show that differences between connected and nonconnected firms are marginally more
important when political links are stronger. Greater differences are found when companies are
connected through owners (rather than directors), through close relationships, or through a min-
ister (rather than a member of parliament). Differences are also greater when the firm operates
in countries with higher degrees of corruption. This last result, in particular, complements the
findings in Faccio (2006), who shows larger net benefits of connections in more corrupt coun-
tries. This paper indicates some of the channels through which such increase in value may be
realized.

References

Aggarwal, R.K., F. Meschke, and T.Y. Wang, 2007, “Corporate Political Contributions: Investment or
Agency,” University of Minnesota Working Paper.
Agrawal, A. and C.R. Knoeber, 2001, “Do Some Outside Directors Play a Political Role?” Journal of Law
and Economics 44, 179-198.
Backman, M., 1999, Asian Eclipse: Exposing the Dark Side of Business in Asia, Singapore, Wiley.
Bortolotti, B., M. Fantini, and D. Siniscalco, 2001, “Privatization: Politics, Institutions and Financial Mar-
kets,” Emerging Markets Review 2, 109-137.
Boubakri, N., J.-C. Cosset, and W. Saffar, 2009, “Politically Connected Firms: An International Event
Study,” American University of Sharjah and HEC Montreal Working Paper.
Calvi, F. and J.-M. Meurice, 1999, Série Noire au Crédit Lyonnais, Paris, Albin Michel Editions.
Campbell, J., 1996, “Understanding Risk and Return,” Journal of Political Economy 104, 298-345.
Chiu, M.M. and S.W. Joh, 2004, “Loans to Distressed Firms: Political Connections, Related Lending,
Business Group Affiliations, and Bank Governance,” Chinese University of Hong Kong Working Paper.
Choi, S.-D., I. Lee, and W. Megginson, 2010, “Do Privatization IPOs Outperform in the Long-Run?”
Financial Management 39, 153-185.
Claessens, S., S. Djankov, and L.H.P. Lang, 2000, “The Separation of Ownership and Control in East Asian
Corporations,” Journal of Financial Economics 58, 81-112.
Cull, R. and L.C. Xu, 2005, “Institutions, Ownership and Finance: The Determinants of Profit Reinvestment
among Chinese Firms,” Journal of Financial Economics 77, 117-146.
De Soto, H., 1989, The Other Path: The Invisible Revolution in the Third Worlds, New York, Harper & Row.
Dewenter, K.L. and P.H. Malatesta, 1997, “Public Offerings of State-Owned and Privately-Owned Enter-
prises: An International Comparison,” Journal of Finance 52, 1659-1679.
Dinç, I.S., 2004, “Politicians and Banks: Political Influences on Government-Owned Banks in Emerging
Countries,” Journal of Financial Economics 77, 453-479.
Faccio, M., 2006, “Politically Connected Firms,” American Economic Review 96, 369-386.

13
Empirical evidence of higher borrowing by politically connected firms is also reported in Chiu and Joh (2004), Cull
and Xu (2005), and Johnson and Mitton (2003).
926 Financial Management ! Autumn 2010
Faccio, M. and L.H.P. Lang, 2002, “The Ultimate Ownership of Western European Corporations,” Journal
of Financial Economics 65, 365-395.
Faccio, M., R.W. Masulis, and J.J. McConnell, 2006, “Political Connections and Corporate Bailouts,” Journal
of Finance 61, 2597-2635.
Ferguson, T. and H.-J. Voth, 2008, “Betting on Hitler—The Value of Poltical Connections in Nazi Germany,”
Quarterly Journal of Economics 123, 101-137.
Fisman, R., 2001, “Estimating the Value of Political Connections,” American Economic Review 91, 1095-
1102.
Fisman, D., R. Fisman, J. Galef, and R. Khurana, 2006, “Estimating the Value of Connections to Vice-
President Cheney,” Columbia University Working Paper.
Gay, P.-A. and C. Monnot, 1999, François Pinault Milliardaire ou les Secrets d’Une Incroyable Fortune,
Paris, Editions Balland.
Goldman, E., J. Rocholl, and J. So, 2008, “Political Connections and the Allocation of Procurement Con-
tracts,” Indiana University Working Paper.
Goldman, E., J. Rocholl, and J. So, 2009, “Do Political Connected Boards Affect Firm Value?” Review of
Financial Studies 22, 2331-2360.
Gomez, E.T. and K.S. Jomo, 1997, Malaysia’s Political Economy: Politics, Patronage and Profits, Cam-
bridge, UK, Cambridge University Press.
Gulen, H., M. Cooper, and A. Ovtchinnikov, 2010, “Corporate Political Contributions and Stock Returns,”
Journal of Finance 65, 687–724.
Hutchcroft, P.D., 1998, Booty Capitalism: The Politics of Banking in the Philippines, Ithaca, NY, Cornell
University Press.
Johnson, S. and T. Mitton, 2003, “Cronyism and Capital Controls: Evidence from Malaysia,” Journal of
Financial Economics 67, 351-382.
Khwaja, A.I. and A. Mian, 2005, “Do Lenders Favor Politically Connected Firms? Rent-Seeking in an
Emerging Financial Market,” Quarterly Journal of Economics 120, 1371-1411.
Kroszner, R.S. and T. Stratmann, 1998, “Interest Group Competition and the Organization of Congress:
Theory and Evidence from Financial Services’ Political Action Committees,” American Economic Review
88, 1163-1188.
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.
Ramalho, R., 2003, “The Effects of Anti-Corruption Campaign: Evidence from the 1992 Presidential
Impeachment in Brazil,” M.I.T. Working Paper.
Roberts, B.E., 1990, “A Dead Senator Tells No Lies: Seniority and the Distribution of Federal Benefits,”
American Journal of Political Science 34, 31-58.
Shleifer, A. and R.W. Vishny, 1994, “Politicians and Firms,” Quarterly Journal of Economics 109, 995-1025.
Stigler, G., 1971, “The Theory of Economic Regulation,” Bell Journal of Economics and Management
Science 2, 3-21.
The Economist, 2001, “Berlusconi: An Italian Story,” April 28-May 4.
The Moscow Times, 1996, Yeltsin Gives Bounty of Tax Breaks to Norilsk, August 14.
Faccio ! Differences between Politically Connected and Nonconnected Firms 927
The New York Times, 1998, The Suharto Billions—A Special Report; For Asian Nation’s First Family,
Financial Empire Is in Peril, January 16.
White, H, 1980, “A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for
Heteroskedasticity,” Econometrica 48, 817-838.
Yoshihara, K., 1988, The Rise of Ersatz Capitalism in South-East Asia, Kuala Lampur, Oxford University
Press.