Twin Agency Problems and Debt Management around the World

Abstract

This study examines the impact of twin agency problems (political corruption and minority shareholders’ expropriation) on corporate debt management policies across a large number of countries. Our results show that in more corrupt countries, managers are more likely to shield liquid assets from potential political extraction by maintaining a higher level of leverage. This effect is magnified by the protection of shareholders’ rights. We further show that twin agency problems influence not only the level of debt in capital structures but also other aspects of debt management, including debt maturity, deviation from optimal leverage, capital structure stability, and the leverage speed of adjustments. The findings are robust due to their inclusion of different measures of corruption and a wide range of firm-level and country-level characteristics. Our study has implications for policymakers, as we show that the improvement of the country-level institutional environment and, particularly, addressing corruption can lead to more effective debt management by firms, ultimately resulting in higher firm values.

1. Introduction

Academic research has focused extensively on investigating the consequences of corruption at both the national and organizational levels. Concurrently, numerous studies have examined the financial and economic consequences associated with the protection of minority shareholders’ rights (MSPs). Traditionally, these two areas of research have been investigated in isolation from one another. However, Stulz (2005) argues that these two problems should be studied together as “twin agency problems”, as in most cases, they exist concurrently. Twin agency problems are defined by Stulz (2005) as two simultaneously occurring problems: “the agency problem of corporate insider discretion” and “the agency problem of state ruler discretion”.

In our study, we investigate the impact of corruption and minority shareholders’ rights protection on firms’ debt policies. While the goal of Stulz’s (2005) work is relatively broad, as it suggests that the “twin agency problem” is one of the main factors that limit global financial development, we focus on one aspect that the “twin agency problem” can impact, which is debt financing. Debt is one of the major sources of external financing for companies. The proportion of debt in a capital structure can influence the cost of capital, which, in turn, impacts a firm’s value. Additionally, debt is a source of financial risk, as too much debt can force a company into bankruptcy. Unlike a firm’s business risk, which is often determined by the type of industry a company operates in, the degree of financial leverage (financial risk) can mainly be controlled by management (is mainly under management control). It has been shown that managers adjust levels of debt in response to changes in economic conditions (Hackbarth et al. 2006; Todorov 2020; Gao and Tsusaka 2023), political conditions (Pan et al. 2019; Basharat et al. 2022), and corporate governance (Chang et al. 2014; Gyimah et al. 2021).

Our analysis aims to shed light on variations in debt management across different countries by exploring the effects of corruption and MSP on various aspects of capital structure. Specifically, we examine their impact on deviations from the optimal capital structure, the speed at which firms adjust towards target (optimal) leverage, the maturity of their debt obligations, and the stability of capital structures.

We provide empirical evidence that the level of corruption among government officials has a distinct influence on debt management practices that is separate from the effects of actions taken by corporate insiders. We show that both the level of corruption and MSP significantly affect capital structure decisions, leverage deviations from target levels, the leverage speed of adjustment (SOA), and the utilization of short-term debt. More importantly, our findings remain robust even after accounting for a wide range of firm characteristics, country-level governance and economic indicators, and alternative measures of corruption and MSP indices.

Our study makes a number of notable contributions to the finance literature. First, it contributes to the emerging literature that recognizes the importance of simultaneously investigating the impact of twin agency problems. Second, it presents a comprehensive analysis of how twin agency problems impact various aspects of capital structure, deviations from optimal debt levels, and leverage SOA across a wide array of countries. Last, this research contributes to the existing literature by examining optimal capital structure and leverage SOA from the perspective of the twin agency problem. To the best of our knowledge, we are among the first researchers to examine corporate debt management from the twin agency perspective.

The remainder of the paper is structured as follows: In Section 2, we present a literature review. Section 3 outlines the methodology used in our study, detailing the main variables and data sources. Section 4 presents the empirical findings, and Section 5 provides concluding remarks summarizing our key findings.

2. Literature Review

2.1. Economic Impact of Corruption

The Enterprise Survey by the World Bank states that “Corruption by public officials may present a major administrative and financial burden on firms. Corruption creates an unfavorable business environment by undermining the operational efficiency of firms and raising the costs and risks associated with doing business”.1

The adverse economic effects of corruption are significant at both national and corporate levels. Previous research has extensively documented the detrimental impact of corruption on a country’s economic and financial development. Hodge et al. (2011) illustrated how corruption impedes economic growth by undermining investments, political stability, and human capital development. Moreover, corruption has been found to diminish the return on investment by reducing investment efficiency (Swaleheen 2007; O’Toole and Tarp 2014). Several recent studies provided evidence that corruption impedes the economic development of countries. Examples of such studies include the analysis of the corruption effect on the Kenyan economy by Mose (2021), on MENA economies by Kirşanli (2023), and on the economic growth of BRICS countries by Nizam et al. (2024). Additionally, Farinha and López-de-Foronda (2023) show that even in typically low-corruption regions, such as the European Union, corruption practices and perception have a negative impact on economic growth and innovations.

Some of the channels through which corruption can hinder economic growth include financial markets (Butler et al. 2009), innovations (Ellis et al. 2020; Farinha and López-de-Foronda 2023), and corporate investment (García-Gómez et al. 2022). Butler et al. (2009) investigated the consequences of state corruption on municipal bond issuance, revealing that higher levels of state corruption are associated with increased credit risk and higher bond yields. Overall, their results show a significant impact of corruption on financial markets.

At the firm level, prior research studies the impact of corruption on corporate innovation, firm value, corporate investment, and corporate financing policies. For instance, Ellis et al. (2020) investigated the relation between local political corruption and innovation within the private sector in the US. Their findings indicate that political corruption negatively affects the quality, quantity, and efficiency of firms’ innovation in the US, underscoring the negative impact of corruption on firms’ innovative activities. García-Gómez et al. (2022) documented the negative impact of corruption on corporate investments in Europe.

2.2. Impact of Corruption on Firm’s Corporate Policies

Previous studies have documented the impact of political corruption on firms’ financial policies in the United States. Smith (2016) shows that firms operating in US areas characterized by higher levels of corruption tend to exhibit lower cash holdings and higher debt levels in their capital structure compared to those in areas with lower corruption levels. These findings are consistent with the asset-shielding hypothesis, suggesting that firms may reduce liquidity and increase debt as a strategic measure to mitigate the risk of expropriation by local officials. Consequently, we expect that firms in countries with higher levels of corruption should have a higher proportion of debt in their capital structure, allowing them to minimize the funds susceptible to expropriation.

However, an alternative hypothesis posits that firms may want to maintain financial flexibility to facilitate bribe payments. Several finance studies provide evidence supporting this perspective, suggesting that certain firms may derive benefits from engaging in corrupt practices (e.g., Fisman 2001; Faccio et al. 2006; Duchin and Sosyura 2012). For instance, political connections can help firms gain preferential access to government contracts and financing. This would imply that countries with higher corruption levels may exhibit lower debt levels as firms prioritize financial flexibility to establish or strengthen their political ties.

Several studies have examined the relationship between legal and institutional environments and various aspects of debt management. For example, Öztekin and Flannery (2012) showed that stronger institutions lower the cost of leverage adjustments. Chang et al. (2014) documented that firms with weaker corporate governance adjusted slower towards target leverage. Zeitun and Goaied (2022) found that a low level of managerial ownership was associated with higher leverage SOA in Japan. Godlewski (2020) showed that stronger creditor protection has a positive impact on renegotiating debt maturity.

The results of studies examining the impact of corruption on debt management are somewhat mixed. For example, Fan et al. (2012) demonstrated that firms in more corrupt countries tend to hold more debt, particularly short-term debt, while Singh and Kannadhasan (2020) found a heterogeneous relationship between corruption and capital structure in emerging economies. Additionally, Hassan et al. (2022) showed that firms operating in areas with high political corruption in the US tend to adopt less short-term debt. Furthermore, while Tran et al. (2023) showed that leverage speed of adjustment (SOA) is negatively influenced by local corruption in Vietnam, Baxamusa and Jalal (2014) identified a non-linear relationship between corruption and SOA across multiple countries and showed that corruption increases the cost of debt and equity.

2.3. Twin Agency Problems

While the impact of corruption on corporate financial policies has been increasingly studied, the majority of these studies overlook the existence of the “twin agency problems” as described by Stulz (2005). The twin agency problems “arise because rulers of sovereign states and corporate insiders pursue their own interests at the expense of outside investors” (Stulz 2005, p. 1595). Stulz (2005) argued that a higher degree of government corruption prompts corporate actions aimed at minimizing state expropriation, such as reducing firm transparency. However, these actions also increase opportunities for corporate insiders to expropriate wealth from outside investors. Consequently, concentrated ownership becomes more efficient than dispersed ownership, leading to a “principal–principal” conflict—an agency conflict which arises between majority and minority shareholders. Therefore, corporate financial policies worldwide often reflect not only the level of corruption in the country (potential wealth expropriation by state rulers) but also the severity of the “principal–principal” conflict and the strength of minority shareholders’ rights protection laws and regulations (Ellul 2008; Amin and Liu 2020; Cao et al. 2022). For example, Ellul (2008) showed that family firms in countries with weak minority shareholders’ protection have significantly higher leverage than non-family firms. Amin and Liu (2020) documented a non-linear association between debt financing and control ownership among Singaporean firms. Cao et al. (2022) found that the participation of minority shareholders in on-line voting leads to a reduced level of debt in the capital structure of Chinese firms.

In this study, we investigate the impact of the twin agency problem on capital structure across the globe, considering that a high level of debt can potentially serve as a mitigating mechanism for both twin agency problems. High debt levels enable firms to shield assets from state rulers’ expropriation while reducing incentives for controlling shareholders to maximize private benefits, as they risk losing control of the firm in the event of bankruptcy.

3. Methodology, Variable Definitions, and Data

3.1. Methodology

To evaluate the effect of corruption and MSP on debt management, we estimate the following model:

$${Leverage}_{i,t+1}={\beta \ast TAP Variables}_{i,t }+{\gamma \mathsf{X}}_{i,t }+\mathsf{\delta }{\mathrm{Y}}_{j,t}+\mathsf{\eta }{\mathrm{Z}}_{j,t}+\epsilon$$

(1)

where TAP (Twin Agency Problems) Variables are the corruption and MSP measures described below, β represents a vector of coefficients, ${\mathsf{X}}_{i, t}$ is a vector of observable firm-level characteristics, and γ is a vector of coefficients. ${\mathrm{Y}}_{j,t}$ is a vector of observable country-level economic and corporate governance factors, and $\mathsf{\delta }$ is a vector of coefficients. ${\mathrm{Z}}_{j,t}$ is a vector of cultural country-specific characteristics, and $\mathsf{\eta }$ is a vector of coefficients. We use ordinary least squares (OLS) regression to estimate Equation (1) and include industry and year fixed effects and firm-level clustering. Country fixed effects are included in the models wherever appropriate.

Our main dependent variable is the level of debt, Leverage, which is defined as the sum of short-term and long-term debt scaled by total assets. The main independent variable is Control of Corruption obtained from the World Bank Corporate Governance database. Control of Corruption evaluates how effectively corruption is controlled within a country and measures perceptions of the degree to which public officials use their power for private gain. It reflects how effective institutions, policies, and practices are in preventing corruption and ensuring accountability for wrongdoing. A higher score on the Control of Corruption indicates stronger anti-corruption management and better governance.

In order to account for the twin agency problems, we additionally include a measure of minority shareholders’ protection, which is the Guillen–Capron Shareholder Rights Index from Guillén and Capron (2016). Additionally, we include an ownership concentration measure, as Stulz (2005) showed that it is a crucial factor to consider when analyzing the twin agency problem’s implication for firms’ policies.

To control for firm-level factors, we include relevant firm characteristics such as size (measured as a log of total assets), market-to-book ratio (measured as total assets minus book equity plus market value of equity, scaled by total assets), tangibility (measured as tangible assets, such as property, plants, and equipment, scaled by total assets), and profitability (measured as earnings before interest and taxes scaled by total assets). These firm characteristics have been identified in the prior literature as main determinants of a firm’s capital structure (e.g., by Rajan and Zingales 1995).

To account for country-level characteristics, we include country economic and financial market variables from the World Bank, such as (log of) GDP per capita, inflation (measured by CPI), domestic credit (measured by domestic credit provided scaled by GDP), market capitalization (total stock market value scaled by GDP), and stock market turnover (measured by total stock value traded divided by stock market capitalization). We also include corporate tax rate from the KPMG Global tax rate survey for 2003–2016.

Several prior studies, including those by Öztekin and Flannery (2012); Öztekin (2015); Singh and Kannadhasan (2020); Zeitun and Goaied (2022), showed that a country’s corporate governance characteristics affect debt management. To control for corporate governance factors, we include the following measures from the World Bank Corporate Governance indicators: Political Stability (reflects perceptions of the likelihood of political instability), Rule of Law (measures perception of confidence in and compliance with the laws), Government Effectiveness (reflects perception of the quality of public and civil service, and extend of its independence from political pressures), Regulatory Quality (measures perceptions the government’s ability to introduce and implement policies related to development of private sector), and Voice and Accountability (measures perceptions of the freedom of expression in the country and the ability of a country’s citizens to participate in government selection process).

Finally, a number of studies showed that national culture impacts capital structure across different countries (e.g., Chui et al. 2002; Mogha and Williams 2021; Orlova and Harper 2022). To control for variability of country-level cultural characteristics, we include the following cultural variables based on Hofstede’s cultural index (Hofstede 1980, 1983, 2001; Hofstede et al. 2010): Individualism, Power Distance, Masculinity, Uncertainty Avoidance, Long-term Orientation, and Indulgence2.

3.2. Data

We obtained firm-level financial variables from the Compustat Global database for the 1995–2016 period. We ended our sample in 2016 because data on our key explanatory variable—minority shareholder protection—are only available until that year. We eliminated firms for which the total asset value is not available. Additionally, countries with less than 15 unique firms and missing values for our main variables of interest were eliminated from the sample. To reduce the effect of outliers, all firm-level variables were winsorized at the 1st and 99th percentiles3. We obtained the Control of Corruption measure from the World Bank Corporate Governance database. Guillen–Capron Shareholder Rights Index values were obtained from Guillén and Capron (2016). The ownership concentration measure was from La Porta et al. (2006) and Djankov et al. (2008). Corporate governance and economic and financial information were obtained from the World Bank’s World Development Indicators database. The Hofstede’s cultural dimensions index was used to measure cultural differences among countries.

Table 1. Country-level summary statistics. This table reports summary statistics for all countries in our sample. Column 1 shows the country to which the analyzed firms belong. Column 2 reports the number of unique firms by country. Column 3 reports the median firm size. Columns 4 reports the median leverage ratio. Column 5 shows median values for the World Bank Control of Corruption variable. Column 6 reports the median Minority Shareholders Protection Index. Variables are as defined in the Section 3. The sample consists of all industrial firms in the Compustat file from 1995 to 2016, with non-missing observations on total assets.

Country Name	Unique Firms	Size	Leverage	Control of Corruption (WB)	Minority Shareholders Index
Argentina	61	5.77	0.18	−0.43	5.75
Australia	2249	2.93	0.03	1.97	6.75
Austria	123	6.27	0.23	1.95	4.50
Bangladesh	163	7.47	0.28	−0.91	2.00
Belgium	151	5.84	0.23	1.38	5.00
Brazil	297	6.47	0.26	−0.02	5.50
Bulgaria	49	4.65	0.23	−0.24	5.42
Canada	489	4.47	0.21	1.99	6.75
Chile	138	10.73	0.23	1.45	4.25
China	3221	7.17	0.18	−0.48	6.66
Colombia	37	13.64	0.11	−0.30	5.50
Cyprus	56	4.13	0.24	1.07	5.00
Czech Republic	23	8.36	0.16	0.33	5.25
Denmark	187	6.48	0.21	2.43	3.00
Egypt	131	6.48	0.14	−0.59	5.92
Finland	165	5.40	0.24	2.37	6.00
France	1047	5.30	0.19	1.36	7.00
Germany	985	4.87	0.14	1.86	6.00
Ghana	13	4.97	0.09	−0.07	4.50
Greece	231	4.95	0.32	0.25	5.00
Hong Kong	1185	6.78	0.15	1.81	6.90
India	3165	6.65	0.28	−0.42	5.88
Indonesia	374	13.36	0.28	−0.74	4.25
Italy	336	6.11	0.26	0.31	6.85
Japan	3965	10.55	0.21	1.23	7.00
Jordan	111	2.83	0.12	0.18	5.50
Kazakhstan	15	9.37	0.22	−0.91	8.25
Kenya	31	8.31	0.12	−0.97	3.00
Latvia	26	2.14	0.13	0.18	5.25
Lithuania	29	4.84	0.27	0.24	6.25
Luxembourg	37	6.77	0.23	2.06	6.25
Malaysia	1007	5.37	0.18	0.28	6.25
Mexico	133	9.20	0.23	−0.37	3.25
Netherlands	247	6.09	0.20	2.18	1.75
New Zealand	155	4.53	0.21	2.35	6.75
Nigeria	98	8.77	0.18	−1.13	5.25
Norway	319	6.48	0.22	2.17	5.25
Oman	52	2.70	0.19	0.28	5.16
Pakistan	314	7.74	0.34	−0.93	3.58
Peru	81	5.89	0.21	−0.30	5.00
Philippines	168	8.04	0.15	−0.53	3.15
Poland	549	4.58	0.13	0.49	7.57
Portugal	73	7.00	0.35	1.08	5.75
Russian Federation	175	9.33	0.23	−0.95	7.35
Singapore	775	4.88	0.17	2.21	7.25
Slovenia	27	7.03	0.27	0.89	6.50
South Africa	369	7.05	0.12	0.34	5.42
Spain	175	7.59	0.26	1.06	4.75
Sweden	739	5.57	0.13	2.25	6.25
Switzerland	262	6.02	0.20	2.17	4.35
Thailand	545	7.59	0.24	−0.31	5.50
Turkey	306	5.64	0.18	0.03	5.90
Ukraine	19	7.76	0.10	−1.00	6.50
United Arab Emirates	42	6.87	0.13	1.08	4.00
United Kingdom	2853	3.80	0.13	1.89	6.63
United States	6327	3.31	0.23	1.39	6.25
Vietnam	369	13.19	0.23	−0.56	6.75
Total/median	35,269	6.47	0.21	0.31	5.50

presents the summary statistics for the median value of leverage, firm size, degree of corruption, minority shareholder protection, and ownership concentration across countries. As shown in Table 1, our sample consisted of 35,269 unique firms from 57 countries. The median value of the debt-to-total assets ratio is 0.21. The median value for the Control of Corruption measure is 0.31, with Denmark showing the highest value of 2.43 and Nigeria showing the lowest value of (−1.13)4. Minority Shareholders’ Protection Index (MSPI) values can potentially range from 0 to 10, as the index reflects 10 different legal provisions from countries’ national legislations. The median value for MSPI in our sample is 5.5.

Table 2. Correlation matrix. This table reports the correlation coefficients for leverage, firm characteristics, country-level corruption, and shareholder protections measures. All variables are as defined in the Section 3. The sample consists of all industrial firms in the Compustat Global database from 1995 to 2016.

VARIABLES	Leverage	Control of Corruption (WB)	Minority Shareholders Index	Size	Market-to-Book Ratio	Profitability	Tangibility
Leverage	1
Control of Corruption (WB)	−0.1464	1
Minority Shareholders Index	−0.0799	0.1912	1
Size	0.1445	−0.2348	−0.0137	1
Market-to-Book Ratio	−0.0271	0.1367	0.0412	−0.3449	1
Profitability	−0.0043	−0.0029	−0.0003	0.0014	0.0001	1
Tangibility	0.0027	0.0021	0.0003	−0.0166	0.05	−0.007	1

presents correlation coefficients showing the relations between leverage, corruption, MSPI, ownership concentration, and various firm characteristics. Notably, we found a negative correlation between debt ratio and corruption. To further explore this relation, we next conducted multivariable analysis to investigate the association between leverage and corruption.

4. Empirical Results

4.1. Effect of Corruption on Level of Debt

Our initial empirical analysis focused on assessing the impact of corruption on the level of debt held by firms. The findings are summarized in

Table 3. Impact of Corruption on leverage. This table presents estimates of OLS regression with leverage as a dependent variable. Leverage is short-term debt plus long-term debt, divided by total assets. The main independent variables include corruption, minority shareholder protection, and ownership concentration measures, as defined in the Section 3. The control variables include firm- and country-level economic, institutional, and cultural characteristics. All variables are as defined in the Section 3. The sample consists of all industrial firms in the Compustat Global database from 1995 to 2016. All specifications include industry and year fixed effects. Standard errors (in parentheses) are heteroskedasticity-consistent and clustered at the firm level. Significance levels are indicated as follows: * = 10%, ** = 5%, *** = 1%.

VARIABLES	Leverage	Leverage	Leverage	Leverage	Leverage
	(1)	(2)	(3)	(4)	(5)
Control of Corruption (WB)	−0.0169 ***	−0.0133 ***	−0.0281 ***	−0.0357 ***	−0.0276 ***
	(0.0010)	(0.0010)	(0.0037)	(0.0046)	(0.0053)
Size		0.0073 ***	0.0080 ***	0.0085 ***	0.0078 ***
		(0.0004)	(0.0004)	(0.0005)	(0.0006)
Market-to-book Ratio		0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***
		(0.0000)	(0.0000)	(0.0000)	(0.0000)
Profitability		−0.0000	−0.0000	0.0000	−0.0000 ***
		(0.0000)	(0.0000)	(0.0000)	(0.0000)
Tangibility		0.0000	0.0000	0.0000 **	0.0000 *
		(0.0000)	(0.0000)	(0.0000)	(0.0000)
Government Effectiveness			−0.0256 ***	−0.0057	−0.0427 ***
			(0.0044)	(0.0054)	(0.0067)
Political Stability			−0.0261 ***	−0.0151 ***	−0.0060 *
			(0.0022)	(0.0029)	(0.0032)
Regulatory Quality			0.0056 *	−0.0130 **	−0.0071
			(0.0034)	(0.0063)	(0.0065)
Rule of Law			0.0533 ***	0.0595 ***	0.0513 ***
			(0.0049)	(0.0071)	(0.0074)
Voice and Accountability			0.0116 ***	0.0134 ***	0.0278 ***
			(0.0019)	(0.0024)	(0.0038)
GDP				0.0016	0.0182 ***
				(0.0025)	(0.0032)
Inflation				−0.0005	0.0004
				(0.0004)	(0.0004)
Domestic Credit				−0.0003 ***	−0.0001 **
				(0.0000)	(0.0000)
Tax Rate				0.0008 ***	0.0018 ***
				(0.0003)	(0.0003)
Market Capitalization				0.0000 ***	−0.0000 ***
				(0.0000)	(0.0000)
Stocks traded turnover				−0.0001 ***	−0.0001 ***
				(0.0000)	(0.0000)
Power Distance					0.0003 **
					(0.0001)
Individualism					−0.0003 **
					(0.0001)
Masculinity					−0.0009 ***
					(0.0001)
Uncertainty Avoidance					−0.0007 ***
					(0.0001)
Long-term Orientation					−0.0002
					(0.0001)
Indulgence					−0.0007 ***
					(0.0001)
Observations	362,961	213,751	213,751	130,452	126,243
R-squared	0.0696	0.0805	0.0904	0.0916	0.0985

, which presents the relationship between corruption and capital structure across various specifications. The table includes different specifications: column 1 considers only the corruption variable, column 2 includes corruption along with firm-level variables, column 3 adds country-level governance variables, column 4 incorporates country economic variables, and column 5 includes country-level cultural characteristics.

Across all specifications tested in our analysis, we consistently observed a negative and statistically significant coefficient on the Control of Corruption variable at the 1% level in Table 3. This finding indicates that firms exhibit a propensity to maintain higher levels of debt in countries characterized by weaker Control of Corruption (i.e., higher corruption). Our results are in line with Smith (2016), who found higher levels of debt for firms in higher-corruption areas in the US. This empirical evidence strongly supports the shielding hypothesis, which posits that firms strategically adjust their leverage levels in response to varying degrees of corruption within their operating environments. These results are consistent with firms strategically using higher debt levels as a defense against the risk of asset expropriation in corrupt countries. This approach helps firms to mitigate the impact of corruption on financial stability and protects assets from potential exploitation.

4.2. Effect of Twin Agency Problems on Level of Debt

Next, we investigated the impact of the twin agency problems on capital structure, incorporating MSPI variable into our regression. Additionally, we included the Ownership Concentration variable, given its increased prevalence in environments where twin agency problems are more acute. The findings presented in

Table 4. Twin agency problems’ impact on leverage. This table presents estimates of OLS regression with leverage as a dependent variable. Leverage is the sum of short-term debt and long-term debt, scaled by total assets. The main independent variables include corruption, minority shareholder protection, and ownership concentration measures, as defined in the Section 3. The control variables include firm- and country-level characteristics. All variables are as defined in the Section 3. The sample consists of 35,269 unique firms from the Compustat Global database from 1995 to 2016. All models include industry and year fixed effects and clustering at the firm level. Standard errors (in parentheses) are heteroskedasticity-consistent. Significance levels are indicated as follows: * = 10%, ** = 5%, *** = 1%.

VARIABLES	Leverage	Leverage	Leverage	Leverage	Leverage	Leverage
	(1)	(2)	(3)	(4)	(5)	(6)
Control of Corruption (WB)		−0.0500 ***			−0.0265 ***	−0.0414 ***
		(0.0062)			(0.0053)	(0.0061)
Minority Shareholders Index	0.0033 *	0.0048 **		0.0104 ***		0.0112 ***
	(0.0019)	(0.0019)		(0.0021)		(0.0021)
Ownership Concentration			0.0017 ***	0.0022 ***	0.0017 ***	0.0019 ***
			(0.0002)	(0.0002)	(0.0002)	(0.0002)
Size	0.0080 ***	0.0082 ***	0.0088 ***	0.0091 ***	0.0087 ***	0.0091 ***
	(0.0006)	(0.0006)	(0.0006)	(0.0006)	(0.0006)	(0.0006)
Market-to-book Ratio	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Profitability	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Tangibility	0.0000 *	0.0000 *	0.0000 *	0.0000 *	0.0000 *	0.0000 *
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Government Effectiveness	−0.0526 ***	−0.0310 ***	−0.0490 ***	−0.0416 ***	−0.0375 ***	−0.0250 ***
	(0.0065)	(0.0071)	(0.0062)	(0.0064)	(0.0069)	(0.0071)
Political Stability	−0.0085 **	−0.0064 *	−0.0037	−0.0021	−0.0021	−0.0010
	(0.0035)	(0.0034)	(0.0035)	(0.0038)	(0.0035)	(0.0037)
Regulatory Quality	−0.0008	0.0058	−0.0066	0.0005	−0.0044	0.0060
	(0.0071)	(0.0070)	(0.0067)	(0.0072)	(0.0067)	(0.0071)
Rule of Law	0.0245 ***	0.0491 ***	0.0388 ***	0.0186 **	0.0524 ***	0.0394 ***
	(0.0081)	(0.0088)	(0.0074)	(0.0081)	(0.0079)	(0.0087)
Voice and Accountability	0.0366 ***	0.0382 ***	0.0170 ***	0.0227 ***	0.0157 ***	0.0257 ***
	(0.0045)	(0.0045)	(0.0043)	(0.0047)	(0.0043)	(0.0047)
GDP	0.0125 ***	0.0143 ***	0.0069 **	−0.0011	0.0093 ***	0.0018
	(0.0034)	(0.0034)	(0.0032)	(0.0035)	(0.0032)	(0.0035)
Inflation	0.0002	0.0001	0.0003	−0.0003	0.0003	−0.0003
	(0.0004)	(0.0004)	(0.0004)	(0.0005)	(0.0004)	(0.0005)
Domestic Credit	−0.0000	−0.0001 **	0.0000	0.0001 **	0.0000	0.0000
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Tax Rate	0.0013 ***	0.0016 ***	0.0012 ***	0.0010 ***	0.0015 ***	0.0012 ***
	(0.0003)	(0.0003)	(0.0003)	(0.0003)	(0.0003)	(0.0003)
Market Capitalization	−0.0000 ***	−0.0000 ***	−0.0001 ***	−0.0001 ***	−0.0000 ***	−0.0000 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Stocks Traded Turnover	−0.0001 ***	−0.0001 ***	−0.0001 ***	−0.0001 ***	−0.0001 ***	−0.0001 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Power Distance	0.0003 **	−0.0001	0.0004 ***	−0.0001	0.0002	−0.0004 ***
	(0.0001)	(0.0002)	(0.0001)	(0.0001)	(0.0001)	(0.0002)
Individualism	−0.0004 ***	−0.0006 ***	0.0003 **	0.0002	0.0003 **	−0.0000
	(0.0001)	(0.0002)	(0.0001)	(0.0002)	(0.0001)	(0.0002)
Masculinity	−0.0012 ***	−0.0012 ***	−0.0011 ***	−0.0014 ***	−0.0010 ***	−0.0014 ***
	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)
Uncertainty Avoidance	−0.0006 ***	−0.0007 ***	−0.0005 ***	−0.0005 ***	−0.0006 ***	−0.0006 ***
	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)
Long-term Orientation	0.0001	0.0001	0.0001	0.0002 **	0.0001	0.0002 **
	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)
Indulgence	−0.0006 ***	−0.0005 ***	−0.0009 ***	−0.0007 ***	−0.0009 ***	−0.0006 ***
	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0001)
Observations	117,194	117,194	124,285	115,654	124,285	115,654
R-squared	0.0996	0.1010	0.0992	0.1008	0.0997	0.1017

show a positive relation between minority shareholders’ rights protection and the level of debt in capital structures. This relation holds across all specifications, with the coefficient on the MSPI variable consistently demonstrating significance levels of 10% or better. These results imply that stronger minority shareholders’ protection is linked to higher debt levels within the capital structure. The results are consistent with the idea that higher minority shareholder protection reduces the risk of expropriation, boosting investor confidence and facilitating firms’ access to debt financing. This, in turn, leads to higher leverage ratios.

Notably, the coefficients on Control of Corruption variables remained negative and statistically significant across all models. This finding means that country-level control of corruption impacts levels of debt, even after we account for other significant factors. These results confirm the notion that firms in environments characterized by weaker control of corruption tend to opt for higher levels of leverage as a defensive mechanism against potential risks of expropriation.

We proceeded to assess whether the relations between leverage and twin agency problem measures observed in the full sample hold across different subsamples. The results are presented in

Table 5. Twin agency problems’ impact on leverage based on legal origin and economic development. This table presents estimates of OLS regression with leverage as a dependent variable. Leverage is the sum of short-term debt and long-term debt, scaled by total assets. The main independent variables include corruption, minority shareholder protection, and ownership concentration measures, as defined in the Section 3. The control variables include firm- and country-level characteristics. All variables are as defined in the Section 3. The sample consists of 35,269 unique firms from the Compustat Global database from 1995 to 2016. All models include industry and year fixed effects and clustering at the firm level. Standard errors (in parentheses) are heteroskedasticity-consistent. Significance levels are indicated as follows: * = 10%, ** = 5%, *** = 1%.

	Leverage		Leverage
	Common Law = 1	Common Law = 0	Developed = 1	Developed = 0
VARIABLES	(1)	(2)	(3)	(4)
Control of Corruption (WB)	−0.0323 **	−0.0781 ***	−0.0318 ***	0.0090
	(0.0134)	(0.0084)	(0.0113)	(0.0086)
Minority Shareholders Index	0.0298 ***	0.0131 ***	0.0061	0.0090 ***
	(0.0064)	(0.0029)	(0.0059)	(0.0025)
Ownership Concentration	−0.0057 ***	0.0033 ***	0.0030 **	0.0010 ***
	(0.0019)	(0.0004)	(0.0012)	(0.0004)
Size	0.0154 ***	0.0042 ***	0.0066 ***	0.0120 ***
	(0.0011)	(0.0008)	(0.0010)	(0.0009)
Market-to-book Ratio	0.0001 ***	0.0000 ***	0.0001 ***	0.0001 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Profitability	−0.0002	−0.0000 ***	−0.0009 ***	−0.0000
	(0.0002)	(0.0000)	(0.0003)	(0.0000)
Tangibility	0.0000 ***	0.0000	0.0000	0.0000 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)
GDP	−0.0055	−0.0054	−0.0188	−0.0224 ***
	(0.0143)	(0.0042)	(0.0127)	(0.0057)
Inflation	−0.0032 ***	−0.0055 ***	−0.0102 ***	−0.0006
	(0.0007)	(0.0006)	(0.0016)	(0.0005)
Domestic Credit	0.0001	0.0004 ***	0.0004 ***	−0.0001
	(0.0001)	(0.0000)	(0.0001)	(0.0001)
Tax Rate	0.0036 ***	0.0000	−0.0016 *	0.0053 ***
	(0.0009)	(0.0004)	(0.0008)	(0.0006)
Market Capitalization	−0.0000	−0.0000	−0.0003 ***	−0.0000
	(0.0000)	(0.0000)	(0.0001)	(0.0000)
Stocks Traded Turnover	−0.0002 ***	−0.0001 ***	−0.0002 ***	−0.0001 ***
	(0.0001)	(0.0000)	(0.0000)	(0.0000)
Government Effectiveness	0.0177	−0.0004	−0.0117	0.0171 *
	(0.0128)	(0.0099)	(0.0115)	(0.0101)
Political Stability	0.0257 ***	0.0154 ***	−0.0312 ***	0.0026
	(0.0087)	(0.0056)	(0.0090)	(0.0053)
Regulatory Quality	−0.0232	−0.0035	−0.0210	0.0116
	(0.0148)	(0.0098)	(0.0148)	(0.0094)
Rule of Law	0.0376 *	0.0351 ***	0.0474 *	0.0082
	(0.0220)	(0.0110)	(0.0243)	(0.0108)
Voice and Accountability	0.0069	0.0251 ***	−0.0190	−0.0013
	(0.0130)	(0.0079)	(0.0188)	(0.0066)
Power Distance	0.0002	0.0000	−0.0002	−0.0005 *
	(0.0008)	(0.0003)	(0.0008)	(0.0003)
Individualism	−0.0018 **	0.0002	0.0002	0.0003
	(0.0007)	(0.0002)	(0.0004)	(0.0003)
Masculinity	0.0045 ***	−0.0018 ***	−0.0012 ***	−0.0009 ***
	(0.0010)	(0.0001)	(0.0003)	(0.0002)
Uncertainty Avoidance	0.0077 ***	−0.0007 ***	−0.0003	0.0003
	(0.0020)	(0.0002)	(0.0006)	(0.0002)
Long-term Orientation	0.0083 ***	0.0004 **	−0.0009 ***	0.0004 *
	(0.0023)	(0.0002)	(0.0002)	(0.0002)
Indulgence	0.0029 ***	0.0002	−0.0004	−0.0005 ***
	(0.0009)	(0.0002)	(0.0004)	(0.0002)
Observations	52,744	73,241	52,287	73,698
R-squared	0.1430	0.0776	0.1069	0.1153

. When we examined the subsample of common law countries versus other countries separately, we found that the sign and statistical significance of the coefficients associated with Control of Corruption and MSPI remained consistent and unchanged compared to the full sample. This means that a higher level of corruption is associated with a higher debt level in both groups of common law countries and non-common law countries. Comparing developed and developing countries, we observed different patterns. In the subsample of developed countries, there is a statistically significant negative relationship between Control of Corruption and debt level; however, the coefficient on MSPI is insignificant. However, in the subsample of developing countries, the relation between Control of Corruption and leverage loses significance, meaning we did not find that corruption has an impact on leverage in developing countries. However, the MSPI coefficient exhibited a positive and statistically significant relation at the 1% level. This implies that a strong MSP leads to a higher amount of debt in the capital structure in developing countries.

One possible explanation for these findings is that developed countries are characterized by more mature and liquid financial markets, with better access to information and investor activism. In these environments, there is less of a need to use more leverage, as existing well-established legal and regulatory mechanisms help to ensure fair treatment of minority shareholders and mitigate the risk of exploitation by majority shareholders. Additionally, firms in developing countries might have limited access to formal financing sources such as bank loans or capital markets and rely more heavily on informal channels, where corruption may play a less important role.

4.3. Estimation of Optimal Level of Debt and Leverage SOA

Although the analysis of how corruption and protection of minority shareholders’ rights affect capital structure provides insights into variations in debt levels across countries, it does not determine whether these observed debt levels are optimal for individual firms in terms of maximizing value, nor does it offer insights into the stability of the capital structure and debt maturity.

The tradeoff theory of capital structure argues that each firm has an optimal level of debt that maximizes firm value by balancing the costs and benefits of debt financing. However, due to market frictions, deviations from the optimal capital structure may occur. Therefore, firms should aim to return to their optimal debt level as quickly as possible. Since an optimal level of debt is unobservable, the most common approach in the literature is to use the partial adjustment model to estimate the target level of debt and speed of adjustment to this target (SOA)5:

$${Lev}_{i,t+1}-{Lev}_{i,t}=\lambda \left({Lev}_{i,t+1}^{*}-{Lev}_{i,t}\right)+{\delta }_{i,t+1}$$

(2)

$${Lev}_{i,t+1}^{*}={\beta \mathsf{X}}_{i,t }$$

(3)

$${Lev}_{i, t+1}=(\lambda {\beta )\mathsf{X}}_{i,t}+(1-\lambda ){Lev}_{i,t }+{\delta }_{i, t+1}$$

(4)

• λ coefficient is the adjustment speed towards target, i.e., leverage SOA;
• $Lev$* is the firm’s target ratio.

$X_{i, t}$ is a vector of various firm- and country-level determinants of the firm’s optimal level of debt, such as market-to-book ratio, size, profitability, tangibility, industry, and firm fixed effects. $\beta$ represents a vector of respective coefficients. The λ coefficient, leverage SOA, represents the proportion of the gap between actual level of debt and the optimal capital structure that a firm is able to close each year.

Following Öztekin and Flannery (2012); Öztekin (2015), we used a two-step system generalized method of moments (GMM) estimator (Blundell and Bond 1998) to estimate Equation (4) to determine leverage SOA for each country. Next, we used estimates of $\beta$s and $\lambda$ to calculate the optimal level of debt (using Equation (3)), and deviation from the target was estimated by subtracting the actual ratio of debt in the capital structure for each firm-year from the estimated optimal level of debt (Equation (5)):

$${\widehat{DEV}}_{i,t+1}={\widehat{Cash}}_{i,t+1}^{*}-{Cash}_{i,t}$$

(5)

We used our estimates of the leverage SOA and deviation from the optimal capital structure as dependent variables in our main specification (Equation (1)). A higher leverage SOA means that firms return to their optimal level of leverage relatively quickly. Additionally, we examined the impact of twin agency problems on the debt maturity and stability of the capital structure. We created an indicator variable that equals 1 if a firm issues notes payable (short-term debt); otherwise, it is 0. We estimated debt volatility as a standard deviation of the level of debt for each firm. A higher standard deviation of debt level implies a less stable (more volatile) capital structure. We included the notes payable (short-term debt) indicator variable and leverage volatility as dependent variables in Equation (1). The effect of twin agency problems on debt maturity, deviation from target, capital structure stability, and leverage SOA were evaluated by regressing the dependent variables described above on Control of Corruption, MSPI, and ownership concentration variables. In line with our previous specifications, we also included firm- and country-level variables into our models. The specifications were estimated using OLS regression and included industry and year fixed effects, and firm clustering.

Table 6. Twin agency problems’ impact on various aspects of debt management. This table presents the results of the impact of twin agency problems on debt maturity, leverage deviation from optimal level, capital structure stability, and the leverage SOA. Dependent variables used are short-term debt use (column 1), deviation from optimal leverage (column 2), leverage volatility (column 3), and leverage SOA (columns 4–6). The short-term debt use is an indicator variable equal to 1 if a firm uses notes payable in its capital structure. Deviation from optimal leverage is ${\widehat{DEV}}_{i,t+1}={\widehat{Lev}}_{i,t+1}^{*}-{Lev}_{i,t}$, where Lev* is target leverage. In column 4, Leverage SOA is estimated using the BB GMM estimator (Blundell and Bond 1998). The following specification is estimated: ${Lev}_{i, t+1}=(\lambda {\beta )\mathsf{X}}_{i,t}+(1-\lambda ){Lev}_{i,t }+{\delta }_{i, t+1}$, where X_i,t is a vector of observable firm-level determinants of factors of the target capital structure, β is a vector of coefficients, and $\lambda$ coefficient represents the leverage SOA. The main independent variables include corruption, minority shareholder protection, and ownership concentration measures, as defined in the Section 3. The control variables include firm- and country-level characteristics. The sample consists of 35,269 unique firms from the Compustat Global database from 1995 to 2016. The standard errors are heteroskedasticity-consistent. Significance levels are indicated as follows: * = 10%, ** = 5%, *** = 1%.

	Short-Term Debt	Deviation from Optimal Leverage	Leverage Volatility	Leverage SOA	Leverage SOA Under-Levered	Leverage SOA Over-Levered
VARIABLES	(1)	(2)	(3)	(4)	(5)	(6)
Control of Corruption (WB)	−0.1072 ***	−0.0287 ***	−0.0133 ***	0.0009	−0.0008	0.0292 ***
	(0.0122)	(0.0020)	(0.0008)	(0.0022)	(0.0028)	(0.0013)
Minority Shareholders Index	0.0250 ***	0.0316 ***	0.0020 ***	−0.0151 ***	−0.0096 ***	−0.0218 ***
	(0.0041)	(0.0007)	(0.0003)	(0.0005)	(0.0009)	(0.0005)
Ownership Concentration	0.0033 ***	0.0044 ***	0.0003 ***	−0.0044 ***	−0.0026 ***	−0.0052 ***
	(0.0005)	(0.0001)	(0.0000)	(0.0001)	(0.0002)	(0.0001)
Size	0.0573 ***	−0.0054 ***	0.0003 ***	0.0002	−0.0011 ***	0.0015 ***
	(0.0011)	(0.0002)	(0.0001)	(0.0001)	(0.0002)	(0.0001)
Market-to-book Ratio	−0.0000 ***	−0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Profitability	−0.0000 ***	−0.0000	−0.0000	−0.0000 **	0.0000 ***	−0.0000 **
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Tangibility	0.0000 *	0.0000	−0.0000	−0.0000 ***	−0.0000 ***	−0.0000
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Government Effectiveness	0.0483 ***	−0.0376 ***	0.0076 ***	−0.0063 ***	−0.0343 ***	0.0083 ***
	(0.0135)	(0.0021)	(0.0009)	(0.0019)	(0.0027)	(0.0018)
Political Stability	0.0017	0.0050 ***	−0.0019 ***	0.0259 ***	0.0186 ***	0.0331 ***
	(0.0069)	(0.0016)	(0.0007)	(0.0009)	(0.0016)	(0.0011)
Regulatory Quality	−0.0019	−0.0141 ***	−0.0120 ***	0.0619 ***	0.0502 ***	0.0775 ***
	(0.0145)	(0.0021)	(0.0012)	(0.0022)	(0.0039)	(0.0019)
Rule of Law	0.1182 ***	0.0614 ***	−0.0031 **	−0.0568 ***	−0.0382 ***	−0.1088 ***
	(0.0175)	(0.0028)	(0.0014)	(0.0039)	(0.0050)	(0.0034)
Voice and Accountability	−0.1075 ***	−0.0007	0.0306 ***	−0.0329 ***	−0.0216 ***	−0.0325 ***
	(0.0089)	(0.0014)	(0.0005)	(0.0017)	(0.0025)	(0.0019)
GDP	−0.0139 **	−0.0324 ***	0.0022 ***	0.0238 ***	0.0307 ***	0.0164 ***
	(0.0070)	(0.0007)	(0.0004)	(0.0008)	(0.0012)	(0.0007)
Inflation	−0.0038 ***	−0.0012 ***	0.0003 ***	−0.0008 ***	−0.0010 ***	0.0007 ***
	(0.0009)	(0.0002)	(0.0001)	(0.0001)	(0.0002)	(0.0001)
Domestic Credit	0.0005 ***	0.0000 ***	0.0001 ***	−0.0006 ***	−0.0005 ***	−0.0005 ***
	(0.0001)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Tax Rate	0.0020 ***	−0.0010 ***	0.0007 ***	−0.0032 ***	−0.0056 ***	−0.0008 ***
	(0.0006)	(0.0001)	(0.0000)	(0.0001)	(0.0001)	(0.0001)
Market Capitalization	−0.0001 ***	−0.0000 ***	−0.0000 ***	0.0001 ***	0.0001 ***	0.0001 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Stocks Traded Turnover	−0.0001 ***	0.0001 ***	−0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Power Distance	0.0013 ***	−0.0005 ***	−0.0002 ***	0.0002 ***	0.0003 ***	0.0003 ***
	(0.0003)	(0.0001)	(0.0000)	(0.0000)	(0.0001)	(0.0000)
Individualism	0.0024 ***	0.0013 ***	−0.0007 ***	0.0021 ***	0.0022 ***	0.0021 ***
	(0.0003)	(0.0000)	(0.0000)	(0.0000)	(0.0001)	(0.0000)
Masculinity	−0.0024 ***	−0.0017 ***	−0.0001 ***	0.0004 ***	0.0003 ***	0.0004 ***
	(0.0002)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Uncertainty Avoidance	0.0004 **	−0.0000	−0.0008 ***	0.0004 ***	0.0005 ***	−0.0004 ***
	(0.0002)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Long-term Orientation	0.0003	0.0014 ***	−0.0008 ***	−0.0010 ***	−0.0007 ***	−0.0016 ***
	(0.0002)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Indulgence	−0.0021 ***	0.0007 ***	0.0000	−0.0005 ***	−0.0005 ***	−0.0007 ***
	(0.0003)	(0.0001)	(0.0000)	(0.0000)	(0.0001)	(0.0000)
Observations	125,985	111,434	117,570	117,122	37,936	79,186
R-squared	0.2499	0.7978	0.7502	0.8922	0.9215	0.9209

shows the results of our estimates.

The results presented in column 1 of Table 6 show that in countries with higher control of corruption and lower degree of MSP, firms use short-term financing less than in more corrupt countries and countries with higher MSP. The results are statistically significant at the 1% level. This finding is in contrast with the results of Hassan et al.’s (2022) study, which showed that firms in a high political corruption area in the US hold less short-term debt. The difference in results can potentially be due to the fact that Hassan et al.’s (2022) study focused on the US, which has more developed financial markets and an overall low level of corruption. We further observed that firms in more corrupt countries deviate more from the optimal level of debt (column 2), and their capital structure is less stable (column 3). This can potentially lead to a decrease in firm value6. A higher level of MSP and ownership concentration amplify the effect. The coefficients are significant at the 1% level. The results in columns 4–6 of Table 6 indicate that while in the full sample, corruption does not have a statistically significant impact on leverage SOA, a lower level of corruption (or a higher level of Control of Corruption) has a positive statistically significant effect on leverage adjustments for over-levered firms (firms with an excess amount of debt). This finding implies that in high-corruption countries, firms holding too much debt are slow to decrease their leverage towards an optimal level. This is consistent with the idea that firms in more corrupt countries hold higher (than optimal) amounts of debt in order to shield their assets from expropriation. This supports the finding of Tran et al. (2023) that linked a higher level of corruption to a slow leverage SOA in Vietnam. This finding is also in line with previous studies that argued that weaker legal institutions increase the costs of leverage adjustment (Öztekin and Flannery 2012; Chang et al. 2014).

4.4. Alternative Measure of Corruption

In order to verify the robustness of our findings and to mitigate any potential biases arising from the choice of corruption measure, we employed an alternative measure of corruption sourced from the International Risk Country Guide Database. This corruption measure ranges from 0 to 6, where “6” denotes low corruption, and “0” denotes high corruption levels. The results of this analysis are summarized in

Table 7. Twin agency problems’ impact on leverage with alternative measure of corruption. This table presents results of the impact of twin agency problems on debt maturity, leverage deviation from target, leverage stability, and the speed of adjustment. Dependent variables used are short-term debt use (column 1), deviation from optimal leverage (column 2), leverage volatility (column 3), and leverage SOA (columns 4–6). The short-term debt use is an indicator variable equal to 1 if a firm uses notes payable in its capital structure. Deviation from optimal leverage is ${\widehat{DEV}}_{i,t+1}={\widehat{Lev}}_{i,t+1}^{*}-{Lev}_{i,t}$, where Lev* is target leverage. In column 4, leverage SOA is estimated using the BB GMM estimator (Blundell and Bond 1998). The following specification is estimated: ${Lev}_{i, t+1}=(\lambda {\beta )\mathsf{X}}_{i,t}+(1-\lambda ){Lev}_{i,t }+{\delta }_{i, t+1}$, where X_i,t is a vector of observable firm-level determinants of factors of the target capital structure, β is a vector of coefficients, and $\lambda$ coefficient represents the leverage SOA. The main independent variables include corruption, minority shareholder protection, and ownership concentration measures, as defined in the Section 3. The control variables include firm- and country-level characteristics. The sample consists of 35,269 unique firms from the Compustat Global database from 1995 to 2016. The standard errors are heteroskedasticity-consistent. Significance levels are indicated as follows: * = 10%, ** = 5%, *** = 1%.

	Leverage	Short-Term Debt	Deviation from Optimal Leverage	Leverage Volatility	Leverage SOA	Leverage SOA Under-Levered	Leverage SOA Over-Levered
VARIABLES	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Corruption Measure from IRCG	−0.0008 ***	−0.0034 ***	−0.0012 ***	−0.0002 ***	0.0004 ***	0.0004 ***	0.0006 ***
	(0.0001)	(0.0002)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Minority Shareholders Index	0.0079 ***	0.0181 ***	0.0294 ***	0.0015 ***	−0.0146 ***	−0.0090 ***	−0.0200 ***
	(0.0021)	(0.0041)	(0.0006)	(0.0003)	(0.0004)	(0.0009)	(0.0005)
Ownership Concentration	0.0019 ***	0.0039 ***	0.0045 ***	0.0003 ***	−0.0044 ***	−0.0026 ***	−0.0053 ***
	(0.0002)	(0.0005)	(0.0001)	(0.0000)	(0.0001)	(0.0002)	(0.0001)
Size	0.0084 ***	0.0580 ***	−0.0050 ***	0.0004 ***	0.0001	−0.0014 ***	0.0014 ***
	(0.0006)	(0.0011)	(0.0002)	(0.0001)	(0.0001)	(0.0002)	(0.0001)
Market-to-book Ratio	0.0001 ***	−0.0000 ***	−0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Profitability	−0.0000 ***	−0.0000 ***	−0.0000	−0.0000	−0.0000 ***	0.0000 ***	−0.0000 **
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Tangibility	0.0000 ***	0.0000 *	0.0000	−0.0000	−0.0000 ***	−0.0000 ***	−0.0000
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Government Effectiveness	−0.0162 **	0.0410 ***	−0.0355 ***	0.0034 ***	−0.0104 ***	−0.0397 ***	0.0124 ***
	(0.0063)	(0.0126)	(0.0016)	(0.0008)	(0.0017)	(0.0022)	(0.0017)
Political Stability	0.0019	0.0216 ***	0.0144 ***	−0.0021 ***	0.0229 ***	0.0172 ***	0.0294 ***
	(0.0037)	(0.0071)	(0.0017)	(0.0007)	(0.0009)	(0.0016)	(0.0012)
Regulatory Quality	−0.0130 *	−0.0173	−0.0200 ***	−0.0124 ***	0.0631 ***	0.0495 ***	0.0811 ***
	(0.0070)	(0.0145)	(0.0021)	(0.0012)	(0.0023)	(0.0039)	(0.0018)
Rule of Law	0.0192 **	0.1022 ***	0.0579 ***	−0.0079 ***	−0.0604 ***	−0.0406 ***	−0.0980 ***
	(0.0079)	(0.0165)	(0.0023)	(0.0013)	(0.0037)	(0.0049)	(0.0034)
Voice and Accountability	0.0224 ***	−0.1143 ***	−0.0013	0.0298 ***	−0.0329 ***	−0.0224 ***	−0.0317 ***
	(0.0045)	(0.0088)	(0.0013)	(0.0006)	(0.0017)	(0.0026)	(0.0019)
GDP	−0.0026	−0.0319 ***	−0.0376 ***	0.0010 **	0.0251 ***	0.0313 ***	0.0198 ***
	(0.0035)	(0.0070)	(0.0008)	(0.0004)	(0.0007)	(0.0012)	(0.0007)
Inflation	−0.0016 ***	−0.0022 **	−0.0005 ***	0.0005 ***	−0.0010 ***	−0.0011 ***	0.0001
	(0.0004)	(0.0009)	(0.0001)	(0.0001)	(0.0001)	(0.0002)	(0.0001)
Domestic Credit	0.0000	0.0004 ***	0.0001 ***	0.0001 ***	−0.0006 ***	−0.0005 ***	−0.0005 ***
	(0.0000)	(0.0001)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Tax Rate	0.0019 ***	0.0027 ***	−0.0008 ***	0.0007 ***	−0.0033 ***	−0.0058 ***	−0.0009 ***
	(0.0003)	(0.0006)	(0.0001)	(0.0000)	(0.0001)	(0.0001)	(0.0001)
Market Capitalization	−0.0000 ***	−0.0001 ***	−0.0000 ***	−0.0000 ***	0.0001 ***	0.0001 ***	0.0001 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Stocks Traded Turnover	−0.0001 ***	−0.0002 ***	0.0001 ***	−0.0000 ***	0.0000 ***	0.0001 ***	0.0001 ***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Power Distance	−0.0006 ***	0.0010 ***	−0.0006 ***	−0.0002 ***	0.0003 ***	0.0004 ***	0.0002 ***
	(0.0002)	(0.0003)	(0.0001)	(0.0000)	(0.0000)	(0.0001)	(0.0000)
Individualism	0.0001	0.0031 ***	0.0015 ***	−0.0006 ***	0.0021 ***	0.0021 ***	0.0019 ***
	(0.0002)	(0.0003)	(0.0000)	(0.0000)	(0.0000)	(0.0001)	(0.0000)
Masculinity	−0.0014 ***	−0.0028 ***	−0.0018 ***	−0.0001 ***	0.0004 ***	0.0004 ***	0.0004 ***
	(0.0001)	(0.0002)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Uncertainty Avoidance	−0.0006 ***	0.0005 ***	−0.0001 **	−0.0008 ***	0.0004 ***	0.0005 ***	−0.0004 ***
	(0.0001)	(0.0002)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Long-term Orientation	0.0003 ***	0.0009 ***	0.0015 ***	−0.0007 ***	−0.0011 ***	−0.0008 ***	−0.0016 ***
	(0.0001)	(0.0002)	(0.0000)	(0.0000)	(0.0000)	(0.0001)	(0.0000)
Indulgence	−0.0006 ***	−0.0020 ***	0.0008 ***	−0.0000	−0.0005 ***	−0.0005 ***	−0.0007 ***
	(0.0001)	(0.0003)	(0.0001)	(0.0000)	(0.0000)	(0.0001)	(0.0000)
Observations	125,985	125,985	111,434	117,570	117,122	37,936	79,186
R-squared	0.0993	0.2515	0.8129	0.7473	0.8931	0.9223	0.9209

. Remarkably, all estimates generated from this alternative corruption measure are consistent with our previous findings. This consistency across the measures confirms the significant association between corruption levels and firms’ debt management decisions.

5. Conclusions

This study examined the relation between corruption and corporate debt policies, with a focus on its differential impact across different countries. Specifically, we investigated the impact of twin agency problems, namely political corruption and minority shareholders’ expropriation, on leverage decisions.

First, we examined how corruption variables affect the level of debt. Our findings indicate that in more corrupt countries, management tends to shield liquid assets from potential political extraction by maintaining higher levels of leverage. Moreover, we observed that higher levels of minority shareholder protection (MSP) are further associated with higher levels of debt.

Subsequently, we explored whether the impact of twin agency problems varies based on legal origin and degree of economic development. Interestingly, we discovered that the impact of corruption on leverage is statistically significant solely in developed countries, while MSP and ownership concentration have a stronger impact on capital structure in developing countries. Next, we studied the impact of corruption and MSP on several dimensions of debt management, including the stability of capital structure, deviation from the optimal level of debt, and leverage speed of adjustment. The results show that twin agency problems indeed have a statistically significant effect on debt policies, including the degree of leverage, debt maturity, stability of capital structure, deviation from target, and leverage speed of adjustment. In summary, we demonstrated the overall significance of twin agency problems for firm financing policies in the international context.

While our analysis included a relatively large sample of firms and countries, it is important to acknowledge that some countries and firms were excluded due to limited or missing data. Thus, the results of the analysis should be interpreted accordingly. While we included a wide variety of control variables in our analysis based on previous research findings, we cannot exclude the possibility that there are some additional factors influencing leverage that have not been identified. Additionally, the findings represent aggregate results for all countries and firms included in our sample and are not necessarily representative of the relationship for any specific country. Future research can focus on country-specific and/or region-specific investigations.