IMPACT OF CORRUPTION ON MNE’S SUBSIDIARY LEVEL OF CASH HOLDINGS

(1)

1

IMPACT OF CORRUPTION ON

MNE’S SUBSIDIARY LEVEL OF

CASH HOLDINGS

by

Wout S. Janssen

S2140160

A thesis submitted in partial fulfilment of the requirements for the degree of

MSc. International Financial Management

University of Groningen

June 2017

(2)

2 Master’s Thesis IFM (EBM022A20)

Date: 09/06/2017. Words: 10422

MSc. International Financial Management

University of Groningen, Faculty of Economics and Business Student number: S2140160

W.S.Janssen@student.rug.nl

Supervisor: Dr. P.P.M. (Peter) Smid University of Groningen

(3)

3

Preface

This master’s thesis encompasses a research carried out from February to June 2017 as partial fulfilment of the requirements for the degree of MSc. International Financial Management from the University of Groningen. The aim of the research was to investigate the impact of corruption on a MNE’s financial policy. Conducting this research was a great learning experience. During this research I got entrained in the different aspects of conducting an academic research. I would like to thank all the ones involved during my research. Especially, I would like to thank Peter Smid for his supervision, feedback, time and effort. I would also like to thank my supervising coach at PwC the Netherlands, Jasper Scholten, and Bas Rebel, for their interest, support and feedback during my research. Peter as well as Jasper and Bas were exceptionally helpful and their feedback and advice helped me to finish this research successfully.

(4)

4

Abstract

This research is an empirical investigation whether the level of corruption in a specific host-country has a significant impact on the level of cash holdings of a multinational enterprise (MNE). I find that there exists a positive relationship between host-country corruption and the cash holdings of foreign subsidiaries affiliated with a MNE headquartered in the United States. An increase of the control of corruption estimate by 1 (a decrease in the perceived corruption) leads to a decrease in cash-to-total assets ratio by 10.5%.

(5)

5

List of Tables

Table 1. Sample information: number of subsidiaries per country... 19

Table 2. Sample information: number of subsidiaries per industry. ... 20

Table 3. Sample information: number of subsidiaries per corruption category ... 20

Table 4. Sample information: Average cash ratio per country. ... 21

Table 5. Descriptive statistics ... 23

Table 6. Correlation table ... 23

Table 7. Regressions on the determinants of corporate cash holdings ... 25

(7)

7

1. Introduction

Corruption is omnipresent and significant all over the world (Shleifer and Vishny, 1993; Jain, 2001). In order to indicate the severity; Transparency International (TI) rates countries on a scale from 0 to 100 every year. This rating stands for highly corrupt and very clean respectively. TI concludes that over two-third of the 176 countries are still suffering from corruption (see figure 1). Moreover, Goel et al. (2016) mention that corruption can result in losses on companies as well as societies as a whole: an estimated cost of 5 percent of the global gross domestic product arise due to corruption. On top of that, survey evidence is confirming the pervasiveness of corruption around the world by pointing out that about 20% of companies have to deal with bribe requests from government employees (Smith, 2016).

What is the impact of these findings on corruption, and to what extent can host-country corruption explain the variations in cash holdings from multinational enterprises’ subsidiaries across different countries? The objective of this study is to investigate whether the level of corruption in a specific host country impacts the level of cash holdings of a subsidiary belonging to a multinational enterprise (MNE) with an ultimate owner located in the US.

(8)

8 Several studies provide evidence that country-specific factors might be more important in explaining a company’s capital structure than firm-specific factors. For example, Fan et al. (2012) found that a country’s institutional environment can have an effect on the way companies are financed. They state that the country where a company is located is more important in determining the company’s capital structure than the industry where the company is doing business in. Other studies (e.g., Doidge et al., 2007; Drobetz and Gruninger, 2007) strengthen this statement by finding that governance variables on country level are of great importance in interpreting corporate liquidity. They argue that these country-level variables could be even more important than the specific variables on firm-level in explaining the liquidity of a company, which is in line with the view of Fan et al.

Corruption is an example of such a country-level variable, so it could be one of the important determinants of corporate cash holdings. It is important to clearly define corruption, since the definition of corruption will determine what will be measured. Firstly, Shleifer and Vishny (1993) define government corruption as selling government property by government employees for personal gain. Another, more broad, definition of corruption is given by Jain (2001), pp.3: “the activities in which public officials, bureaucrats, legislators and politicians use powers delegated to them by the public to further their own economic interests at the expense of the common good”. This research will follow the latter definition. In other words, firm-level corruption is not part of this study. Generally, there exists two opposite views on corruption which are both empirically supported. One view is positive about corruption and argues that corruption is efficient and allows firms to avoid bureaucratic rules (Smith, 2016). It could even raise economic growth (Mauro, 1995). The second view sees corruption as inefficient which results in losses for companies (and society). It can be seen as a maleficent type of taxation (Shleifer and Vishny, 1993). Thus, in this second view, companies are looking at bribes as a sort of tax that they do not want to pay.

(9)

9 This existing literature suggest that companies hold more cash in their overseas subsidiaries because of higher tax costs in their home-country. Despite of this findings, this insight with regard to tax does not explain all variation in MNE cash holdings (Pinkowitz et al., 2016). Moreover, other determinants of corporate cash holdings are extensively discussed within the current literature (see e.g. Opler et al. 1999; Ozkan and Ozkan, 2004; Ferreira and Vilela, 2004; Kim et al, 1998; Maheshwari and Rao, 2017; García-Teruel and Martínez-Solano, 2008). This study investigates another potential country-level determinant for cash holdings in MNEs’ foreign subsidiaries: corruption. There is a rising concern and an increasing number of studies on the impact of corruption. Only Smith (2016) found that there is a negative relationship between corruption and corporate cash-holdings within the United States (US) by investigating different areas. Current studies that examine the impact of corruption, like Smith (2016), are based on individual countries or companies. However, to my best knowledge, the impact of host-country corruption on the cash holdings of a MNE have not been studied before. Beuselinck and Du (2017) also suggest that future research should examine how corporate governance quality of host-countries influences MNE’s subsidiary cash holdings. This leads to the research question of this study: What is the impact of host-country corruption on the

cash holdings of a MNE’s subsidiary in that host country?

This firm-level analysis will result in a better insight into the impact of corruption on the cash holdings of MNEs. As a practical contribution, the results can be helpful for managers at controlling subsidiaries evaluating the cash holding dimension of the corporate financial policies and to encounter possible early warning signs of corruption. A literature review is carried out first, where after a hypothesis is set up. Thereafter, to be able to test the hypothesis, a regression model is developed and control variables are selected. Lastly the data is collected using the Orbis database and the hypothesis is tested by carrying out log-linear OLS regressions in EViews. Specifically, the study empirically examines firm-level and country-level data from 2015 of 1989 subsidiaries located in 24 different countries, from 553 MNEs headquartered in the United States.

(10)

10

2. Literature review

In this section a review of the existing literature is carried out on the impact of corruption, specifically on cash holdings.

Several studies find that corruption has a negative impact in general. For example, Goel et al. (2016) states that corruption can lead to more income inequality, dilutes citizen’s faith in political services and can lead to a decrease in investments in education, the environment and health care. Furthermore, corruption is acting as a significant deterrent to development and growth (Jain, 2001). Moreover, a politically corrupt environment reduces the economic efficiency (Smith, 2016). This is enforced by the empirical findings of Bellavite Pellegrini et al. (2017), who suggests that the relation between a country corruption index and the total annual return of the stocks of the listed industrial companies is negative. This suggestion is consistent with Donadelli et al. (2014), who concluded that firms that do business in countries with a high level of corruption have lower returns than firms that are operating in countries with a low level of corruption. Other studies find empirical evidence for a negative impact of host-country corruption on foreign direct investment (FDI) (Wei, 2000; Cuervo-Cazurra, 2006). This evidence is enforced by Habib and Zurawicki (2002), who finds that corruption causes operational inefficiencies and thereby weakens the incentive to invest of foreign investors. On the other hand, Williams & Kedir (2016) found that corruption does not reduce growth rates and (annual) sales at all.

2.1. Cash holdings

(11)

11 Moreover, MNEs can increase the cash holdings of subsidiaries to keep up the profits of these subsidiaries and, through this, invest more in the foreign country where the specific subsidiary is located (Arena and Kutner, 2015). On the contrary, they state that MNEs also can recur foreign profits to the home country to increase domestic cash holdings and corporate pay-outs. However, multinational companies tend to hold their cash in the host-country to fence off taxes in regard of repatriation (Foley et al., 2007).Naiwei (2011) finds that a firm’s liquidity is lower in countries with higher control of corruption. Another finding of Naiwei (2011) is that holding too much cash can lower firm value in countries with low control of corruption in comparison to other countries. Regarding cash holdings, Smith (2016) carried out a study within the US and finds that firms that are headquartered in more corrupt areas hold less cash than firms located in less corrupt areas and that this effect is stronger for firms that are less geographically diversified. In contrast with this finding of Smith, Pinkowitz et al. (2006) found that there is a positive relationship between cash holdings and corruption. Dudley and Zhang (2016) find evidence that there exist a positive relationship between the cash holdings of a company and the level of societal trust in a country, especially in countries where there are weak institutions. Gonenc and Seifert (2016) find that when the level of corruption is low, managers decide to hold less cash. This is explained by the managers’ assumption that future funding is easily available in low corruption countries. To summarize, they argue there exists a positive relationship between corporate cash holdings and corruption.

2.3. Hypothesis

Based on the above reviewed literature in the foregoing section and the corresponding research question drawn in the introduction, this study will test the following hypothesis using ordinary least squares (OLS) regressions:

H1: Subsidiaries of US MNEs hold more cash in more corrupt host-countries than in cleaner

(12)

12

3. Data and methodology

The methodology of the study will be discussed in this section. This consists of the regression model, the measurement of the variables and the (gathering of) data.

3.1. Regression model

To test the hypotheses, OLS regressions will be carried out to examine the impact of corruption on cash holdings while controlling for other cash holding determinants. In order to empirically investigate this, it is done using the following regression model:

Cash = α + β1(Corruption) + β2(Distance) + β3(Size) + β4(Leverage) + β5(NWC) (1)

+ β6(Cash flow) + β7(Risk-taking) + β8(Inflation) + β9(Tax) + β10(Subsidiaries)+ ε

The coefficient of main interest of this study is the corruption coefficient, β1. As mentioned in

the literature review and the hypothesis in section two, the variance in the ratio of cash holdings to total assets of foreign subsidiaries might be explained by host-country corruption. Likewise, subsidiary cash holdings is the dependent variable within this model and corruption is the independent variable in the regression. Logically, there are other determinants of cash holdings where this study needs to control for. These controls are geographical distance, size, leverage, risk-taking, net working capital, cash flows, inflation, tax and the number of subsidiaries. The next section will evaluate why this study chose the selected controls.

3.2. Variables

(13)

13 As mentioned before, MNE’s subsidiary cash holdings is the dependent variable in this study. Some studies use the logarithm of cash to assets net of cash, which avoids some of the econometric problems in scaling the dependent and independent variables by assets and corrects for the large outlier firms that carry substantial cash ratios (e.g., Microsoft, Google). Forthcoming from the literature, there are different ways of measuring cash holdings, for example the ratio of the sum of cash and cash equivalents to total assets minus cash and cash equivalents (Sha and Sha, 2016; Ramírez and Tadesse, 2009).

Another measure is the ratio of cash and short-term investments to total assets (Fernandes and Gonenc, 2016). Smith (2016) uses the ratio of cash and cash equivalents divided by total assets. Likewise, Foley et al. (2007) use the ratio of cash to total assets to measure cash holdings. Moreover, Maheshwari and Rao (2017) as well as Beuselinck and Du (2017), Naiwei (2011) and Opler (1999) measure cash holdings as the ratio of cash to total assets.

Following the majority of the existing literature on cash holdings, this study will also use the cash-to-total assets ratio to measure the cash holdings of a company. Cash is measured as the cash at bank and in hand of the subsidiary. To measure total assets, the book value is used.

Corruption is the independent variable in this study. Chen et al. (2015) highlight three different indicators to measure corruption at a country level. The main one is the Transparency International's (TI) Corruption Perception Index for measuring corruption. This index is also used in other studies, for example in Wei (2000), Aidt (2009) and Mo (2001). However, Lambsdorff (2008) argues that an increase or decrease of the index score is not explained by the corruption of the country per se. He states that it can also change due to the methodology that TI uses or an increase/decrease in the amount of surveys. To reduce the bias discussed above, Chen et al. (2015) divides the index by the mean of indices from all countries. Another indicator is the 'control of corruption' (COC), this is a sub-index of the World Bank's Worldwide Governance Indicators and is frequently used in the literature, see e.g. Chen et al. (2015), Cuervo-Cazurra (2006) and Naiwei (2011).

(14)

14 This study will use COC because it uses polls of experts as well as surveys of citizens and people in business in the country. Besides this, it has the widest coverage of all measures of corruption available (Kaufmann et al. 2010). Countries are arranged in two different ways: based on their rank, and based on their estimate of governance. The rank is the percentile rank among all countries, which ranges from 0 (lowest) to 100 (highest). The higher the rank, the lower the corruption. The estimate of governance is a score given based on a country’s governance performance. This score ranges from approximately -2.5 to 2.5. A high index score indicates a high control of corruption, which means less perception of corruption. A low index score indicates that there is a higher perception of corruption, because of a lower control of corruption. This study will use the estimate of governance score to measure corruption. Moreover, robustness tests are carried out by using the rank proxy of corruption.

Next to the dependent variable and the independent variable, the study includes control variables to obtain more robust results. The control variables that this study use are obtained from existing literature.

First of all, Ramírez and Tadesse (2009) found that firm multinationalism moderates the effects that culture has on the cash holdings of a company. However, the more a company is diversified on geographical as well as industrial level, the less cash a company holds on average (Fernandes and Gonenc, 2016).

(15)

15 To overcome a bias because of the size of the US, we measure the distance in kilometres from New York to the capital of the country where the subsidiary is located. This is calculated using the great circle formula.

Most MNEs are relatively large companies. Due to the fact that they are more diversified and have larger resource bases, they have a lower probability of financial distress. Therefore, it can be expected that larger companies hold less cash than smaller companies (Rajan and Zingales, 1995). In accordance with this, Opler et al. (1999) came to the conclusion that small firms have higher cash holdings, because larger firms are more able to access capital markets and are therefore able to hold less cash. Consequently, the Size of a MNE’s subsidiary is used as a control variable. García-Teruel and Martínez-Solano (2008) measure size as the natural logarithm of total assets. This is in line with Opler et al. (1999), who use the natural logarithm of the book value of total assets. Following both studies, Beuselinck and Du (2017) uses the natural logarithm of the total assets of a subsidiary. Likewise, Maheshwari and Rao (2017) measure size using the natural logarithm of the book value of total assets. Naiwei (2011) also measure size as the book value of total assets. Following this existing literature, this study will also proxy size by the natural logarithm of the book value of total assets of an ultimate owners’ subsidiary. Total assets are measured by the book value of the fixed assets plus the book value of the current assets.

Ozkan and Ozkan (2004) find that one of the significant determinants of corporate cash holdings is leverage. Moreover, Ferreira and Vilela (2004) as well as Maheshwari and Rao (2017) concluded that there exists a negative relationship between corporate cash holdings and leverage. Naiwei (2011) states that there is an existing negative relationship between cash holdings and leverage. His study also suggests that leverage is considered to be an important determinant of the liquidity of a firm, because it could diminish the agency problem.

Consequently, this study will use Leverage as a second control variable. To measure leverage,

(16)

16 Beuselinck and Du (2017) measure leverage by dividing the total debt by the total assets of a subsidiary and find that this has a significant impact on the level of cash holdings of a subsidiary. Likewise, Naiwei (2011) measure leverage by the ratio of total debt to total assets. This study follows the latter measurement method of leverage as well: the total debt to total assets ratio. The total debt is measured as the book value of short term debt plus the book value of long term debt and the total assets are also measured by their book value.

It is also important to take into consideration other liquid assets that could function as a substitution for cash, and measure those specific assets. Therefore this study includes Net

working capital as a control variable, in accordance with Opler et al. (1999), Ozkan and

Ozkan (2004), Naiwei (2011) and Beuselinck and Du (2017). Net working capital is measured by Beuselinck and Du (2017) as inventories plus receivables minus payables to total assets. However, Naiwei (2011) proxy net working capital by using total current assets minus cash and total current liabilities. This is in line with the way that Opler et al. (1999) measure net working capital. Consequently, this study will measure net working capital by subtracting both cash and the current liabilities from the total current assets. This is then divided by the total assets. For the current assets as well as the current liabilities this study uses the book value.

(17)

17 In the view of the agency theory, more risk-averse CEOs want to reduce firm risk and therefore tend to hold more cash. This is sometimes even at the cost of the shareholder value. In line with the negative relationship that the agency theory predicts, Tong (2010) finds that companies hold less cash when they have a CEO which takes on more risk. Palazzo (2012) shows that a firm’s riskiness is positively related to the cash holdings of that particular firm. Opler et al. (1999) find that US companies with more riskier cash flows tend to hold more cash relative to companies with less risky cash flows. Following this literature, the study also includes Risk-taking as a control variable. John et al. (2008) argues that returns become more volatile when companies have more riskier corporate operations, and that the risk-taking of a MNE can be measured as the country-adjusted volatility of firm level earnings before interest and tax (EBIT). This method is consistent with Bruno and Shin (2014), who state that firms which take more risk have a tendency of having more volatile returns and that when a MNE’s EBIT is compared with the average of a specific country, this can be measured. This research will measure a MNE’s level of risk-taking likewise the previous mentioned two studies. Specifically, risk-taking is measured by dividing the sum of the standard deviations of the EBIT of the subsidiary from 2010 to 2015 by the sum of standard deviations of the country average EBIT from 2010 to 2015.

Pinkowitz et al. (2006) presents that inflation is negatively related to corporate cash holdings, since companies want to avoid a decrease of the value of the cash due to high inflation. Other studies confirm this and find that inflation can, indeed, explain a decrease in cash that is held by companies, due to a decrease in purchasing power. (e.g. Huang et al.,2013; Wang et al.,2014). Thus, differences in purchasing power because of inflation could impact the cash holdings of companies at the macro-level. For this reason, Inflation is a control variable in this study. Inflation can be measured by the annual inflation rate, which can be found in the World Bank Database. This database measures inflation by the annual growth rate of the GDP implicit deflator, which shows the rate of price change in the economy as a whole. The GDP implicit deflator is the ratio of GDP in current local currency to GDP in constant local currency. This study measures inflation by using the realised inflation rates of 2015.

(18)

18 This study uses the corporate tax rates from 2015 that are applied in the host countries. The corporate tax rates are obtained from the ‘corporate tax rates table’, which is developed by KPMG1.

Finally, Beuselinck and Du (2017) suggest that monitoring subsidiaries will be more costly when the MNE has more subsidiaries. They argue that this could induce an increase in subsidiary cash holdings, because MNEs decide to let the subsidiaries do the decision-making themselves. By investigating their own suggestion, they found that subsidiary cash holdings and the total number of MNE subsidiaries are positively related. Therefore, Subsidiaries is included in the regression model as control variable. This is measured by the total number of companies in the corporate group of a subsidiary. The corporate group of a subsidiary includes all ultimately owned subsidiaries of the subsidiaries’ ultimate owner. Thus, the number of companies in the corporate group of a subsidiary is the same for subsidiaries that are owned by the same ultimate owner.

3.3. Data

The dataset used in this study is gathered by employing the Orbis database, published by Bureau van Dijk. By using this database, it is possible to find companies based on ownership structures, financial information and governance indicators. This database enables gathering financial –and ownership data of consolidated MNEs and the unconsolidated accounts of their subsidiaries abroad (Beuselinck and Du, 2017).

This is important because immediate repatriation of cash, or cash pooling, can have an exceptional impact on the level of cash holdings of a specific company. This study starts by determining which firms from the dataset are appropriate to include in the sample. All the data from the sample are from 2015. The only exception on this is the data for the risk-taking variable. To measure risk-taking, EBIT data from 2010 to 2015 is used. The data is gathered on all foreign subsidiaries, owned by an industrial and listed ultimate owner2_{located in the United}

States (US). The choice for only including subsidiaries whereof the listed ultimate owners are located in the United States in the sample, is to avoid a home-country bias. The US is ranked as the 18th most clean country of the world.

1 https://home.kpmg.com/xx/en/home/services/tax/tax-tools-and-resources/tax-rates-online/corporate-tax-rates-table.html

(19)

19 The Orbis database makes use of a so called consolidation code, which indicates what type of financial statement is available. There are two broad categories, namely consolidated statements and unconsolidated statements. Subsidiaries with consolidated financial statements have consolidation codes C1 or C2, and subsidiaries with unconsolidated financial statements are assigned to consolidation codes U1 or U2. Remaining companies have consolidation codes LF or NF, which indicates that there is only limited financial data available in case of code LF and no financial data available in case of code NF. To make the subsidiaries more comparable by only have unconsolidated subsidiary accounts in the sample and overcome cash pooling or repatriation biases, it is decided to exclude all subsidiaries with consolidation codes that are not U1 or U2. The following steps are taken out to make the data more appropriate. Firstly, some companies in specific industries are deleted. To start with, financial firms (SIC 6000-6999) are excluded because, among other things, they have required liquidity levels (Smith, 2016). Moreover, subsidiaries of utilities (SIC 4900-4999) are and quasi-regulated industries (SIC 4000-4499) are excluded because problems with regard to comparing will arise due to specific balance sheets used within this industries. Second, subsidiaries that are located in the home country, the US, are excluded. Third, all subsidiaries that (indirectly) hold a subsidiary or branch themselves are deleted to overcome a bias due to cash pooling. Next, host-countries where less than 5 subsidiaries are located are deleted from the sample. Fourth, companies with missing data and outliers are deleted. This resulted in a final sample of 1989 subsidiaries, across 24 different countries3, of 553 MNEs. More information about the resulting sample is displayed in table 1, table 2, table 3 and table 4.

Country Number of subsidiaries Country Number of subsidiaries Country Number of subsidiaries

Austria 18 Greece 21 Philippines 13

Belgium 54 Croatia 16 Poland 115

Bulgaria 16 Hungary 42 Romania 92

Czech Republic 57 Ireland 52 Serbia 9

Germany 68 Italy 338 Sweden 88

Estonia 17 Korea 26 Slovenia 26

Finland 31 Norway 81 Slovakia 43

France 342 New Zealand 30 UK 394

Total 1989

Table 1. Sample information: number of subsidiaries per country.

3_{Austria, Belgium, Bulgary, Croatia, Czech Republic, Estonia, Finland, France, Germany, Greece, Hungary,} Ireland, Italy, Korea, New Zealand, Norway, Philippines, Poland, Romania, Serbia, Slovakia, Slovenia, Sweden, United Kingdom.

(20)

20 Table 1 displays that France, Italy and the United Kingdom have the most subsidiaries in the sample. The least subsidiaries are located in Serbia. The average number of subsidiaries per country is 83.

Division SIC Code

Number of subsidiaries

Agriculture, Forestry and Fishing 0100-0999 4

Mining 1000-1499 5

Construction 1500-1799 41

Not used 1800-1999 0

Manufacturing 2000-3999 549

Transportation, Communications, Electric, Gas and Sanitary service 4000-4999 53

Wholesale Trade 5000-5199 634

Retail Trade 5200-5999 155

Finance, Insurance and Real Estate 6000-6799 0

Services 7000-8999 547

Public Administration 9100-9729 1

Non-classifiable 9900-9999 0

Total 1989

Table 2. Sample information: number of subsidiaries per industry.

Based on table 2, it can be concluded that the most of the subsidiaries are manufacturing companies, wholesale trade companies, retail trade and service companies.

Corruption estimate Number of subsidiaries Corruption rank Number of subsidiaries

-2.5 to -1.5 0 0-20 0 -1.5 to -0.5 0 20-40 0 -0.5 to 0.5 673 40-60 489 0.5 to 1.5 518 60-80 325 1.5 to 2.5 798 80-100 1175 Total 1989 Total 1989

Table 3. Sample information: number of subsidiaries per corruption category. The corruption rank is the percentile rank

among all countries, which ranges from 0 (lowest) to 100 (highest). The corruption estimate of governance is a score given based on a country’s governance performance. This score ranges from approximately -2.5 to 2.5.A high index score points to a high control of corruption, which means less perception of corruption. A low index score state that there is a higher perception of corruption, because of a lower control of corruption.

(21)

21

Country

Average

cash ratio Country

Average

cash ratio Country

Average cash ratio

Austria .082 Greece .290 Philippines .290

Belgium .186 Croatia .440 Poland .211

Bulgaria .269 Hungary .187 Romania .224

Czech Republic .217 Ireland .146 Serbia .314

Germany .149 Italy .165 Sweden .207

Estonia .423 Korea .258 Slovenia .282

Finland .210 Norway .246 Slovakia .258

France .157 New Zealand .238 United Kingdom .208

Total .198

Table 4. Average cash ratio per country. The average cash ratio is the average of all companies’ cash ratios per country, the

(22)

22

4. Results

This section discusses the descriptive statistics, correlations and results of the hypotheses tests.

4.1. Descriptive statistics and correlations

The descriptive statistics for the sample of this study are illustrated in table 5. On average, 19.8 percent of the total assets of a company is cash. The rank of corruption has an average of 79.8 and ranges from 41.8 to 100, which indicates that there is a wide spread of corruption ranks. The theoretical range of the corruption governance estimate is -2.5 to 2.5. In the sample, it ranges from -0.43 to 2.29. From the 210 countries worldwide, 132 countries fall within this range of corruption. This confirms that the sample represents a wide spread of different host-country corruption levels. Table 6 displays the correlation coefficients. These coefficients suggest that companies hold less cash when they take more risk, are larger and if they are located in a more corrupt host-country. Moreover, the table shows that there are no variables that are strongly correlated as there are no correlation coefficients that are high enough to indicate the presence of serial correlation.

(23)

23

Variable Mean Std. Dev. Median Minimum Maximum

Cash 0.198 0.236 0.096 0.000 1 Corruption (1) 1.073 0.857 1.280 -0.430 2.291 Corruption (2) 79.785 16.588 87.980 41.827 100 Distance 6546.669 1419.938 6327 5121 14420 Size 15.995 1.627 16.032 9.427 22.202 Leverage 0.496 0.256 0.491 0.000 1 NWC 0.206 0.333 0.192 -0.826 2.059 Cash flows 0.123 0.222 0.091 -3.417 4.286 Risk-taking 0.744 2.614 0.216 0.001 77.145 Inflation 0.878 1.165 0.625 -2.304 4.879 Tax 25.106 6.605 22 10 33.99 Subsidiaries 407.82 421.166 293 3 3622

Table 5. Descriptive statistics. N=1989. Cash is defined as the cash-to-total assets ratio. Corruption (1) is the corruption

estimate, which is defined as the score given based on a country’s governance performance. This score can range from approximately -2.5 to 2.5. A high index score indicates a high control of corruption, which means less perception of corruption. A low index score indicates that there is a higher perception of corruption, because of a lower control of corruption. Corruption (2) is the corruption rank, which is the percentile rank among all countries, which ranges from 0 (high corruption) to 100 (low corruption). Distance is defined as the number of kilometres from New York to the capital of the host-country using the great circle formula. Size is defined as the natural logarithm of total assets. Leverage is defined as the total debt to total assets ratio. NWC is defined as the current assets minus current liabilities and cash divided by total assets. Cash flows is defined as the EBITDA minus interest and taxes, scaled by total assets. Risk-taking is defined as the country-adjusted volatility of firm level earnings before interest and tax. Inflation is defined as the realised inflation rates of 2015 in the host countries. Tax is defined as the corporate tax rates from 2015 that are applied in the host countries.

Subsidiaries is defined as the total number of companies in the corporate group of a subsidiary. The corporate group of a

subsidiary includes all ultimately owned subsidiaries of the subsidiaries’ ultimate owner.

1 2 3 4 5 6 1. Cash 1 2. Corruption (1) -0.006 1 3. Corruption (2) -0.017 0.989 1 4. Distance 0.059 -0.327 -0.375 1 5. Size -0.221 0.119 0.096 0.051 1 6. Leverage -0.190 -0.079 -0.074 -0.029 -0.009 1 7. NWC -0.387 0.002 -0.011 -0.005 0.093 -0.469 8. Cash flows -0.005 0.006 0.015 0.062 -0.093 -0.137 9. Risk-taking -0.027 -0.013 -0.017 -0.014 0.343 0.016 10. Inflation -0.045 -0.106 -0.087 -0.015 0.060 -0.010 11. Tax -0.098 -0.109 -0.056 0.035 -0.157 0.184 12. Subsidiaries -0.041 -0.077 -0.075 0.031 0.050 -0.001

Table 6. Correlation table. N=1989. Cash is defined as the cash-to-total assets ratio. Corruption (1) is the corruption estimate,

(24)

24 7 8 9 10 11 12 1. Cash 2. Corruption (1) 3. Corruption (2) 4. Distance 5. Size 6. Leverage 7. NWC 1 8. Cash flows 0.003 1 9. Risk-taking 0.004 -0.025 1 10. Inflation 0.047 0.064 0.026 1 11. Tax -0.048 -0.020 -0.038 -0.408 1 12. Subsidiaries -0.003 0.070 0.005 -0.008 0.005 1 Table 6. Continued. 4.2. Hypotheses tests

(25)

25 (1) (2) (3) (4) Corruption -0.020 -0.105* -0.007** -0.010* Distance 0.000*** Size -0.370*** -0.354*** -0.353*** -0.029*** Leverage -2.367*** -2.424*** -2.436*** -0.428*** NWC -2.87*** -2.910*** -2.909*** -0.417*** Cash flows -0.333* -0.267 -0.262 -0.086*** Risk-taking 0.017 0.012 0.012 0.004** Inflation -0.176*** -0.187*** -0.186*** -0.010** Tax -0.062*** -0.062*** -0.062*** -0.003*** Subsidiaries -0.001*** -0.001*** -0.001*** -0.000* N 1989 1989 1989 1989 Adj. R-squared .248 .240 .241 .369

Table 7. Regressions on the determinants of corporate cash holdings. N=1989. The dependent variable is Cash, which is

defined as the log of the cash-to-total assets ratio in (1), (2) and (3). In (4), Cash is defined as the cash-to-total assets ratio.

Corruption in (1) and (2) is the corruption estimate, which is defined as the score given based on a country’s governance

performance. This score can range from approximately -2.5 to 2.5.A high index score points to a high control of corruption, which means less perception of corruption. A low index score state that there is a higher perception of corruption, because of a lower control of corruption. Corruption in (3) is defined as the percentile rank among all countries, which ranges from 0 (lowest) to 100 (highest). The higher the rank, the lower the corruption. Distance is defined as the number of kilometres from New York to the capital of the host-country using the great circle formula. Size is defined as the natural logarithm of total assets. Leverage is defined as the total debt to total assets ratio. NWC is defined as the current assets minus current liabilities and cash. Cash flows is defined as the EBITDA minus interest and taxes, scaled by total assets. Risk-taking is defined as the country-adjusted volatility of firm level earnings before interest and tax. Inflation is defined as the realised inflation rates of 2015 in the host countries. Tax is defined as the corporate tax rates from 2015 that are applied in the host countries. Subsidiaries is defined as the total number of companies in the corporate group of a subsidiary. The corporate group of a subsidiary includes all ultimately owned subsidiaries of the subsidiaries’ ultimate owner. *** = p-value < 0.01, ** = p-value < 0.05, * = p-value < 0.1

Model (1) examines the relationship between host-country corruption and the level of subsidiary cash holdings, while controlling for other determinants of corporate cash holdings. As the results show, distance is significant on the 1% significance level (p < 0.01). However, the coefficient is zero.

(26)

26 Interpreting the corruption coefficient leads to the following impact of corruption on cash holdings: If the control of corruption estimate of a host-country increases (decreases) by 1, the cash ratio of subsidiaries located there decrease (increase) by 10.5%. The higher the control of corruption estimate, the lower the perceived corruption. Practically, it means that there exists a positive relationship between the cash ratio of a subsidiary and the corruption in the host-country where the subsidiary is located. In other words: ceteris paribus, the subsidiaries located in the most corrupt country in our sample have 28.6 percent4_{higher cash ratios than}

subsidiaries located in the most clean host-countries. The hypothesis, that there exists a positive relationship between cash holdings and corruption, is supported. While interpreting the other country-level determinants of cash holdings included in the model the study finds that if the tax rate in a host-country increases (decreases) by 1, the cash ratio of subsidiaries located in the specific host-country decrease (increase) by 6.2%. An increase (decrease) of 1 subsidiary in the corporate group of the MNE results in a decrease (increase) of 0.1% in cash ratio. Size is negatively related to the cash ratio of subsidiaries in the sample; an increase of 1% of the subsidiary’s size leads to a 0.35% decrease in the cash ratio. Leverage and NWC are also negatively related to cash ratio. An increase of 0.1 in leverage or NWC results in a decrease of 2.4% or 2.9% in cash ratio respectively.

Model (3) displays the results of the robustness test carrying out the same OLS regression as in model (2), but using the other proxy for corruption: the corruption rank. Corruption gets significant on a 5% level (p < 0.05) with a p-value of 0.018. The corruption coefficient changes from -0.105 to -0.007. The coefficients of the control variables remain having the same values. This seems valid because of the high correlation coefficient of 0.989 between the two proxies. The corruption coefficient in model (3) suggests that an increase in corruption rank by 1 leads to a decrease in cash ratio of 0.7%. An increase in rank means that the host-country gets less corrupt, meaning that the host-country corruption and the cash ratio of subsidiaries in that country are positively related. The difference in corruption coefficient using the different proxies can be explained by the fact that the range of the estimate is from -2.5 to 2.5 and the ranking ranges from 0 to 100. This results suggests that an increase of 15 ranks is the same as an increase of the corruption estimate by 1.

(27)

27 Model (4) is a robustness test. It is the same regression as in model (2) but the cash-to-total assets ratio is used as dependent variable, instead of the logarithm of the cash-to-total assets ratio.

The negative coefficient of the control of corruption estimate is significant on a 10% significance level (p < 0.1) with a p-value of 0.059. This result confirms a positive relationship between corruption and a MNE’s subsidiary cash holdings.

(28)

28

5. Conclusion

This section consists of the conclusion of the study, a discussion, the limitations of the study and some possible opportunities for further research.

5.1. Conclusion

(29)

29 The finding of this study, that there exists a positive relationship between host-country corruption and subsidiary cash holdings, makes sense when combining the findings of the studies of Cuervo-Cazurra (2006), Ramírez and Tadesse (2009) and Arena and Kutner (2015). Another aim of this study was to investigate whether the findings of Smith (2016) are also present on an international level. Smith finds that companies in more corrupt US areas shield their assets for corrupt practices by decreasing their level of cash. This study shows that this negative relationship is not present when the scope of the study is extended to an international level.

Besides corruption, the study controls for other possible determinants of cash holdings. The study finds that a negative relationship between the cash holdings of a subsidiary and its size, leverage, net working capital and the number of companies in the corporate group of the MNE exists. Next to that, the study shows that there is a negative relationship between subsidiary cash holdings and the country-level controls: corporate tax and inflation. These relationships are consistent with the findings in the existing literature on this topic.

The findings of this study can lead to important implications for managers and treasurers of MNEs. Beuselinck and Du (2017) suggest that holding cash in foreign subsidiaries has an advantage as well as a disadvantage. The advantage is that subsidiaries are stimulated to be innovative and that they can make use of local (business) opportunities fast. The disadvantage is related to this study and suggests that cash can be used illegitimate or ineffective. Beuselinck and Du (2017, pp 111) state that this happens more often when a MNE does business in a host-country where “the country-level corporate governance mechanisms are underdeveloped”. We can assume that this is the case in corrupt countries most of the time.

(30)

30 Assuming that the survey findings about the visibility of bank accounts and cash is realistic, the findings of this study regarding the impact of corruption on cash holdings can have a practical impact on managers and treasurers within a MNE.

Because this study finds that subsidiaries in more corrupt host-countries hold more cash, they are able to exploit corrupt practice more quickly in those countries.

Area

Percentage of bank accounts visible by central treasury on a daily basis

Percentage of total cash visible by central treasury on a daily basis

United States 83,4% 90,7% Western Europe 77,6% 86,9% South America 77,5% 95,0% Africa 80,1% 92,1% Asia-Pacific 59,6% 73,5% Middle East 92,5% 95,0% Eastern Europe 65,0% 100,0% Total 77,6% 87,3%

Table 8. Bank account –and cash visibility of companies, outcome from PwC Global Treasury Benchmark Survey 2016.

There are more opportunities for corrupt practices in corrupt countries and the subsidiaries located there have the cash already at hand. Cash is necessary for subsidiaries to exploit growth opportunities in the host-country, but this also encompasses a risk of the expropriation of cash. So it is possible that the almost 10% of cash that is not visible, and held at overseas subsidiaries in corrupt host-countries, is used by the subsidiaries for opportunities that arise due to corrupt practices. If this is the case without the headquarter knowing it, because of the invisibility of the cash, it could become known sometime that this company is involved in corrupt practices. Reputational damage is a possible negative consequence that could emerge from this. Thus, the findings of this study could encourage international managers to control and monitor subsidiaries in corrupt host countries better. This could be done by, for example, assigning an expatriate CEO or treasurer to safeguard the MNE’s cash (Beuselinck and Du, 2017).

5.2. Limitations and further research

(31)

31 This is due to the fact that there is not much (correct) financial data available of the very corrupt host-countries. Because of this limitations, the findings of this study may not be generalized blindly.

Further research could investigate a larger dataset by using more home countries to see whether this results are generalizable over a larger amount of countries.

(32)

32

6. Bibliography

Aidt, T.S. (2009). Corruption, institutions, and economic development. Oxford Review of

Economic Policy, 25:2, 271–291.

Arena, M. P., Kutner, G. W. (2015). Territorial Tax System Reform and Corporate Financial Policies. Review Of Financial Studies, 28:8, 2250-2280

Bellavite Pellegrini, C., Sergi, B.S., Sironi, E. (2017). Stock returns, Productivity, and Corruption in Eight European Fast-Emerging Markets. Thunderbird International Business

Review, 59:1, 15-22.

Beuselinck, C., Du, Y. (2017). Determinants of Cash Holdings in Multinational Corporation's Foreign Subsidiaries: US Subsidiaries in China. Corporate Governance: An International

Review, 25:2, 100-115

Boubaker, S., Derouiche, I., Lasfer, M. (2015). Geographic location, excess control rights, and cash holdings. International Review Of Financial Analysis, 42, 24-37

Bruno, V., Shin, H. S. (2014). Globalization of corporate risk taking. Journal Of International

Business Studies, 45:7, 800-820.

Chen, M., Jeon, B.N., Wang. R., Wu, J. (2015). Corruption and bank risk-taking: Evidence from emerging economies. Emerging Markets Review, 24, 122-146

Cuervo-Cazurra, A. (2006). Who cares about corruption? Journal of International Business

Studies, 37:6, 807-822.

Dittmar, A., Mahrt-Smith, J., Servaes, H. (2003). International Corporate Governance and Corporate Cash Holdings. Journal Of Financial & Quantitative Analysis, 38:1, 111-133.

(33)

33 Donadelli, M., Fasan, M., Magnanelli, B.V. (2014). The Agency Problem, Financial Performance and Corruption: Country, Industry and Firm Level Perspectives. European

Management Review, 11, 259-272.

Drobetz, W., Gruninger, M. C. (2007). Corporate cash holdings: Evidence from Switzerland.

Financial Markets and Portfolio Management, 21, 293–324.

Dudley, E., Zhang, N. (2016). Trust and corporate cash holdings. Journal Of Corporate

Finance, 41, 363-387

Fan, J. P., Rui, O. M., Zhao, M. (2008). Public governance and corporate finance: Evidence from corruption cases. Journal Of Comparative Economics, 36:3 , 343-364

Fan, J.P.H., Titman, S., Twite, G. (2012). An international Comparison of Capital Structure and Debt Maturity Choices. Journal of Financial and Quantitative Analysis, 47:1, 23-56.

Ferreira, M. A., Vilela, A. S. (2004). Why Do Firms Hold Cash? Evidence from EMU Countries. European Financial Management, 10:2, 295-319

Fernandes, N., Gonenc, H. (2016). Multinationals and cash holdings. Journal Of Corporate

Finance, 39, 139-154.

Foley, C.F., Hartzell, J., Titman, S., Twite, G. (2007). Why do firms hold so much cash? A tax-based explanation. Journal of Financial Economics, 86:3, 579-607.

García-Teruel, P. J., Martínez-Solano, P. (2008). On the Determinants of SME Cash Holdings: Evidence from Spain. Journal Of Business Finance & Accounting, 35:1-2, 127-149

Goel, R.K., Mazhar, U., Nelson, M.A. (2016). Corruption Across Government Occupations: Cross-National Survey Evidence. Journal of International Development, 28, 1220-1234.

(34)

34 Habib, M., Zurawicki, L. (2002). Corruption and Foreign Direct Investment. Journal Of

International Business Studies, 33:2, 291-307.

Huang, Y., Elkinawy, S., Jain, P. K. (2013). Investor protection and cash holdings: Evidence from US cross-listing, Journal of Banking and Finance, 37, 937–951.

Jain, A.K. (2001). Corruption: A Review. Journal of Economic Surveys, 15:1, 71-123.

John, K., Litov, L., Yeung, B. (2008). Corporate Governance and Risk-Taking, Journal of

Finance, 63, 1679-1728

Kim, C. S., Mauer, D., Sherman, A. E. (1998). The determinants of corporate liquidity: Theory and evidence, Journal of Financial and Quantitative Analysis, 33, 335–359.

Maheshwari, Y., Rao, K.T.V. (2017). Determinants of Corporate Cash Holdings, Global

Business Review, 18:2, 1-12

Malshe, A., Agarwal, M. K. (2015). From Finance to Marketing: The Impact of Financial Leverage on Customer Satisfaction. Journal Of Marketing, 79:5, 21-38

Mauro, P. (1995) Corruption and growth. Quarterly Journal of Economics, 110:3, 681-714.

Naiwei, C. (2011). Securities Laws, Control of Corruption, and Corporate Liquidity: International Evidence. Corporate Governance: An International Review, 19:1, 3-24

Palazzo, B. (2012). Cash holdings, risk, and expected returns. Journal of Financial Economics, 104, 162-185

Pinkowitz, L., Stulz, R., Williamson, R. (2006). Does the Contribution of Corporate Cash Holdings and Dividends to Firm Value Depend on Governance? A Cross-country Analysis.

(35)

35 Pinkowitz, L., Stulz, R. M., Williamson, R. (2016). Do U.S. Firms Hold More Cash than Foreign Firms Do?. Review Of Financial Studies, 29:2, 309-348

Opler, T., Pinkowitz, L., Stulz, R., Williamson, R. (1999). The determinants and implications of corporate cash holdings. Journal of Financial Economics, 52:1, 3-46

Ozkan, A., Ozkan, N. (2004). Corporate cash holdings: An empirical investigation of UK companies. Journal Of Banking & Finance, 28:9, 2103-2134

Rajan, R.G., Zingales, L. (1995). What do we know about capital structure? Some evidence from international data. Journal of Finance, 50:5, 1421–1460

Ramírez, A., Tadesse, S. (2009). Corporate cash holdings, uncertainty avoidance, and the multinationality of firms. International Business Review, 18:4, 387-403

Shleifer, A., Vishny, R. (1993). Corruption. Quarterly Journal of Economics, 108, 599–617.

Smith, J. D. (2016). US political corruption and firm financial policies. Journal of Financial

Economics, 121, 350-367.

Tong, Z. (2010), CEO Risk Incentives and Corporate Cash Holdings, Journal of Business

Finance & Accounting, 37:9, 1248–1280

Wang, Y., Ji,W., Chen, X., Song, C. 2014. Inflation, operating cycle, and cash holdings,

China Journal of Accounting Research, 7, 263–276.

Wei, S. J. (2000). How taxing is corruption on international investors? The Review of

Economics and Statistics, 82:1, 1-11

(36)

IMPACT OF CORRUPTION ON MNE’S SUBSIDIARY LEVEL OF CASH HOLDINGS