Cash holdings and Multinationality: a European perspective
Abstract
Using data from twelve countries in the European Union over a 13-year period (2002-2015) with 9,707 observations, the effect of multinationality and the crisis on cash holdings is examined in a European setting. Both firm and country characteristics of firms are taken into account. This research contributes to the fields of risk management in the area of cash holdings and multinationality. Findings
suggest that the cash ratio of companies is not significantly related to multinationality or the financial crisis. Moreover, findings show that, when taking determinants of cash holdings into account, Dutch
firms have significantly higher cash holdings than eight out of eleven countries in the sample.
Name: Ruben Hanson
Student number: s2359022
Supervisor: dr. R. M. van Dalen
Second Supervisor: prof. dr. B.W. Lensink
Faculty of Business and Economics, University of Groningen
Study Programme: MSc International Financial Management
2
Table of Content
Introduction p. 3
Literature Review p. 5
Motives for cash holdings p. 5
Firm characteristics p. 6
Country characteristics p. 9
Cash holdings and the crisis p. 10
Cash holding motives and determinants p.11
Data and methodology p. 12
3
Introduction
In the first fiscal quarter of 2017, the cash holdings of Apple were 246.1 billion US dollar. That is the highest cash holding for a non-financial company ever recorded (CNN, 2017). To put that in perspective, the cash holdings of Apple are higher than the GDP of developed countries such as New Zealand and Finland (World Bank, 2017).
The high cash holdings of Apple are striking, but Apple is certainly not the only company that has increased its cash holdings substantially over the last years. According to Moody’s (2016) corporate cash holdings in the United States have more than doubled from 2006 to 2016. This could implicate on the one hand that firms have healthy balance sheets that can cope with adverse shocks. On the other hand it can also be problematic, because these increases in cash holdings can implicate that firms are reluctant to invest their money. This reluctance to invest does not benefit countries that are trying to increase their growth rates and employment. As suggested by Pinkowitz et al. (2016), firms have increased their cash holdings because they are hesitant to invest for the future.
In earlier research, Pinkowitz et al. (2012) focus on the role that multinational corporations have in the increase of cash holdings in the United States. They find that multinational corporations increased their cash holdings from 157 billion dollar in 1998, to 835 billion dollar in 2010. In this period, multinationals increased their cash holdings by 433 percent while they increased their assets by 205 percent. On the other hand, domestic firms increased their cash holdings by 66 percent, while their assets increased by 40 percent. This indicates a strong positive relationship between multinationality and cash holdings. Nonetheless, this cannot be said about all countries, as Fernandes and Gonenç (2016) find that this positive relationship between multinationality and cash holdings only holds in developing countries. In their sample of companies in 58 different countries they even find a negative relationship between multinationality and cash holdings.
4 to 21.6 percent of GDP in 2007. The latest measurement in 2012 shows a difference of 20 percent of GDP (CPB, 2014).
This data presents a clear difference in cash holdings between the Netherlands and the core-EU. One of the factors that could play a role here is multinationality, which provides mixed results in the available literature. On the one hand Pinkowitz et al. (2012) found that in the United States multinationals were responsible for a large part of the increase in cash holdings. On the other hand Fernandes and Gonenç (2016) find a negative relationship between
multinationality and cash holding. These mixed results will be looked into in more detail for the EU sample. This will be done using the following research question:
What is the effect of multinationality on cash holdings in the European Union?
After researching this question in the paper, the findings show no significant relationship between multinationality and cash holdings at any significance level. With regard to the crisis, it is striking that cash holdings decreased sharply when the crisis began, but cash holdings increased to levels even higher than pre-crisis levels in recent years. However, in the
regression analysis there was no significant relationship found between cash holdings and the crisis. When looking at the difference between The Netherlands and other countries in the EU, it is striking to see that Dutch firms have significantly more cash holdings than eight out of eleven countries in the sample.
5
Literature review
The literature review will discuss motives for corporate cash holdings. In the subsequent sections, these four motives will be linked to firm determinants, country determinants of cash holdings and the financial crisis. The final section of the literature review provides a summary of the link between these determinants and the motives for holding cash.
Motives for cash holding
Firms can have different motives to hold cash. According to Bates, Kahle and Stulz (2009) there are four main motives for holding cash: the transaction motive, the precautionary motive, the tax motive and the agency motive. These will be discussed in more detail in this section.
First of all, the precautionary motive argues that firms hold more cash when access to capital markets is costly, for example due to adverse shocks. Firms with riskier cash flows and poor access to external capital hold more cash according to this motive. According to Keynes (1936) the precautionary motive consists of three reasons to hold cash, these are: the ability to pay unforeseen expenses, the ability to take advantage of opportunities related to
advantageous investments and the ability to pay foreseen expenses. In other words, businesses hold cash today to be able to utilize investment opportunities in the future and to be able to pay expenses.
The second motive provided by Bates et al. (2009) is the transaction motive. This motive takes into account the costs that are related to changing a financial asset into cash and using cash for payments. As there are economies of scale with this motive, it is expected that larger firms will hold less cash.
6 Finally, the agency motive is concerned with shareholder and creditor protection by the legal system (Ferreira & Vilela, 2004). In countries with little shareholder and creditor protection entrenched managers are more likely to retain cash than to increase pay-outs to shareholders when the firm has poor investment opportunities. They accumulate cash to be able to
influence the investment decisions of the firm more. In this case bad corporate governance would lead to higher cash holdings.
Firm characteristics and cash holdings
In terms of determinants of cash holdings, a distinction can be made between country
characteristics and firm characteristics. The firm characteristics that can affect cash holdings will be discussed first.
Multinationality
Multinationality can affect cash holdings in two ways. First, based on the transaction motive, large multinationals can have economies of scale with regard to cash management and thus they are expected to hold relatively less cash. This is supported by the findings of Fernandes and Gonenç (2016) who find a negative relationship between cash holdings and
multinationality. However, their results indicate a difference between emerging and developed countries. Cash holdings tend to decrease when foreign sales of multinationals increase in developed markets, while the contrary is the case in emerging markets. Opposed to this reasoning there is the tax motive, which implicates that multinationals hold more cash compared to domestic firms. Bates et al. (2009, p1989) argue that “corporations that would incur tax consequences associated with repatriating foreign earnings hold higher levels of cash”. As multinationals experience the consequences of this repatriation the most, they are expected to hold more cash. This is supported by the findings of Pinkowitz et al. (2012), who find that a large part of the increase in cash holdings in the United States is due to the increase of cash holdings by multinational firms.
7 seems to be no stronger evidence for one of both relationship, the following hypothesis is set up:
Hypothesis 1: Cash holdings in Europe are affected by multinationality of firms.
Substitutes
There are different substitutes of cash holdings that affect the level of cash holdings in a firm. It is expected that these substitutes are negatively related to cash holdings, as there is less need to hold cash when a company is in possession of a substitute (Ferreira & Vilela, 2004). This is in line with the precautionary motive, as substitutes decrease the necessity to hold cash to be able to pay future investments and expenses. Substitutes for which a negative
relationship is expected are net working capital and dividends. Net working capital consists of substitutes for cash, thus assets that can easily be turned into cash. With regard to dividends Opler et al. (1999) mention that firms that pay dividend can raise capital by stopping to pay dividend. Thus for firms that pay dividends to their shareholders, there is less need to hold cash as they are in possession of a substitute.
Another determinant that can be seen as a substitute for holding cash is leverage. When borrowing is seen as a substitute for cash, a negative relationship with cash holdings is expected. However, when leverage of a firm increases, then the firm is increasingly likely to experience financial distress. For this reason these firms increase their cash holdings to decrease the risk of bankruptcy, thus for higher levels of leverage this argument suggests that there is a positive relationship between leverage and cash holdings (Guney, Ozkan & Ozkan, 2007). This can be linked to financially constrained firms, which are discussed next.
Financially constrained firms
8 Firm size
The negative relation between firm size and cash holdings found by Opler et al. (1999) can be due to the fact that smaller firms are often younger and not as well-known as larger firms. Therefore they will need to hold more cash in order to be able to utilize their investment opportunities as they will have less access to capital markets compared to larger firms in the case of market imperfections (Denis & Sibilkov, 2009). This is in line with the precautionary motive. Moreover, it can be argued that there are economies of scale and thus larger firms will hold less cash, which is in line with the transaction motive. (Bates, Kahle and Stulz, 2009).
Investment
According to Haushalter, Klasa and Maxwell (2006) firms with higher cash holdings are more likely to increase their investment compared to other firms in their industry. Based on these results one would expect a positive relationship between cash holdings and investment. Three different types of investment are examined: R&D spending, capital expenditure and
9 Cash flow
The effect of the cash flow on cash holdings can be explained in two ways. On the one hand it can be argued that firms with higher cash flows accumulate more cash and thus have higher cash holdings (Bates et al., 2009). This positive relationship is also found by Opler et al. (1999). On the other hand, firms that have higher cash flows do not need to hold as much cash as firms with lower cash flows as they can replenish their cash holdings more quickly
(Pinkowitz et al, 2016), which is in line with the precautionary motive.
Raising capital
Bates et al. (2009) argue that firms that raise capital have higher cash holdings just after they raised this capital. After this their levels of cash will decrease as they spend the money they raised. This effect of raising capital can be caused both by issuing debt and issuing equity. Using the precautionary motive, it can be argued that firms that raised more capital will have higher levels of cash holdings, as businesses will save this capital for when they need it.
Country characteristics and cash holdings
Next to the firm characteristics that influence cash holdings, the external environment of a firm has an impact on this as well. Therefore the next section will deal with the country characteristics. There are different country characteristics that can be distinguished which can affect corporate cash holdings, such as the quality of institutions and the economic and
financial development of countries. Pinkowitz et al. (2016) distinguish these characteristics as well and argue that institutions have an impact on the type of investment decisions that companies make in a certain country. As the risk of expropriation increases, it become more likely that a company will invest in assets that are harder to expropriate. As this affects the firm characteristics of the companies operating in a country, cash holdings also differ as institutions differ.
Moreover, firms in countries that have a poor investor protection are expected to hold more cash, as they are subject to more agency problems (Dittmar, Mahrt-Smith, 2003). This is confirmed by Pinkowitz et al. (2006), who find that cash that is held by companies in
10 With regard to financial and economic development, Love (2001) finds that companies in countries with more developed financial markets have lower corporate cash holdings. Developed financial markets make it easier for companies to attract money when they need funds, therefore decreasing their need to hold cash in the firm, which is in line with agency theory.
The crisis and cash holdings
The precautionary and transaction motives can explain why the 2007 financial crisis can affect corporate cash holdings. First of all, the precautionary motive states that firms hold more cash when access to capital markets is costly, which can be the case due to adverse shocks such as a financial crisis. Thus, when firms were hit by the crisis, the precautionary motive leads to the expectation that firms decreased their cash holdings, as they were holding cash as a buffer that could be used in case of an adverse shock.
The second motive that affects cash holdings in times of crisis is the transaction motive, which deals with the costs related to converting a noncash financial asset into cash. In a crisis banks do not supply as much loans as usually. This makes borrowing for companies more expensive or even impossible in a crisis. Therefore it is expected that cash holdings decrease for companies during a crisis as they have to use internal funds instead of external funds to finance their operations.
These two motives both expect cash holdings to decrease in times of crisis and these
11 provides either mixed results or a negative relationship between the crisis and cash holdings. As only the negative relationship is supported by the motives, it is expected that there will be a negative relationship between the crisis and cash holdings.
Cash holding motives and determinants
The expected relationships between the motives and the determinants of cash holdings is summarized in table 1. It stands out that only multinationality is expected to have a negative effect on cash holdings based on the transaction motive, while it has a positive effect on cash holdings based on the tax motive. Moreover, the country related determinants (quality of institutions and economic and financial development) and the crisis are expected to have a negative relationship with cash holdings.
Table 1. Expected relationships between motives and determinants
Table 1 describes the expected relationship between the motives and the determinants of cash holdings discussed in the literature review.
Precautionary Transaction Tax Agency
Firm determinants
Multinationality Negative Positive
Substitutes Negative
Financially constrained firms Positive
Firm size Negative Negative
Investment Positive Positive
Cashflow Negative
Raising Capital Positive
Country determinants
Quality of institutions Negative
Economic and financial development Negative Negative
Crisis
12
Data & methodology
Based on the literature above, the effect of multinationality on cash holdings will be
investigated. The following set up will be used to test the hypothesis and to find out what the effect is of multinationality on cash holdings in the European Union.
Model
The following equation is used to answer the research question:
Cash holdingsi,c,t = α + β1*Multinationalityi,c,t + β2*Firm Controli,c,t + β3*Country Control,c,t
+ β4*Crisisi,c,t + 𝜀i,c,t
In which i is a subscript of firms, c is a subscript of countries and t is a subscript of time.
Using this model, the goal is to find the effect of multinationality on cash holdings. This model controls for different firm and country characteristics discussed in the literature review. The following section elaborates on the data composition and the different variables used.
Data
The sample period examined in this research will cover the years 2002 to 2015, this period is chosen to have a large sample of both the pre-crisis and the crisis period. In order to create the sample, countries from the European Union were selected. Next, countries that did not have enough characteristics available were dropped. This resulted in a sample consisting of firms from the Netherlands, Germany, France, Italy, Slovakia, Spain, Portugal, Luxembourg, Ireland, Austria, Belgium and Greece. Firms from these countries were found using the Orbis database, firm codes were then used to download firm data from the Datastream database. Selected firms needed to be headquarters of publicly listed firms. This was done to make sure that every company was present only once in the sample and to make sure that there was enough information. Furthermore, firms that had no data on cash holdings and total assets were omitted from the sample as these variables form the dependent variable. Financial and utility companies are excluded from the sample in line with Opler et al. (1999). They argue that this is necessary because the business of financial firms includes inventories of
13 holdings. Initially, the sample consisted of 1082 firms. After omitting firms from the sample that did not have sufficient data or were not available in the Datastream database, 1040 firms were left in the sample.
The descriptive statistics in table 2 show that there are 9,707 observations of companies over the sample period for the dependent variable cash ratio. The cash ratio has a minimum value of 0 and a maximum value of 0.99 and the reported mean is 0.14.
Table 2. Descriptive statistics
Variable Observations Mean Std. Dev. Min Max
Cash ratio 9.707 0.1362 0.1410 0.0000 0.9889 Multinational 9.707 0.7436 0.4367 0.0000 1.0000 Crisis 9.707 0.6414 0.4796 0.0000 1.0000 Size 9.707 12.8192 2.2260 6.6631 19.7395 Cash flow 9.707 0.0608 0.1304 -2.6627 2.5121 R&D 9.707 10.1461 146.2722 0.0000 7735.2900 Capex 9.707 5.0798 7.2234 0.0000 190.0000 Equity Issuance 9.707 0.0225 0.0964 -0.1116 2.1857 NWC 9.707 0.0164 0.1835 -1.4646 0.7884 Leverage 9.707 0.2392 0.1809 0.0000 0.9960 Industry Volatility 9.707 128489.7000 194352.5000 489.0531 4760697.0000 MTB 9.707 1.6135 3.7509 0.0551 89.7045 Dividend 9.707 0.6384 0.4805 0.0000 1.0000 WGI 9.707 0.0351 1.0148 -2.6873 1.6945 ASDI 9.707 0.3449 0.1069 0.2028 0.7889 RADI 9.707 3.3029 0.8231 2.0000 5.0000 Bank Credit 9.707 96.2092 24.4605 29.8046 172.4112 Turnover 9.707 0.8890 0.5773 0.0015 3.7725 GDP 9.707 39377.0900 9926.5490 11144.4300 110001.1000
14 0 0,02 0,04 0,06 0,08 0,1 0,12 0,14 0,16 0,18
Figure 1. Average cash ratio of firms in the sample
To look into this in more detail, table 3 provides the cash ratio means per country over the sample period of the countries that have ten or more companies in the sample, which leads to the exclusion of Slovakia in the table. The table shows that for every country, the cash ratio falls from 2007 to 2008. Moreover, in every country, apart from Luxembourg, the cash ratio increases again in 2009.
Table 3. Country means of cash holdings for every year in the sample
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Austria .098 .087 .094 .135 .140 .137 .130 .141 .114 .107 .107 .105 .099 .109 Belgium .106 .134 .147 .137 .153 .152 .137 .151 .145 .138 .149 .156 .155 .180 France .142 .142 .147 .157 .153 .146 .143 .156 .158 .156 .153 .165 .180 .186 Germany .158 .166 .167 .181 .180 .174 .161 .170 .169 .162 .168 .173 .176 .169 Greece .076 .075 .070 .059 .062 .070 .066 .078 .077 .069 .070 .071 .074 .081 Ireland .085 .071 .121 .088 .085 .068 .064 .071 .096 .076 .075 .088 .096 .086 Italy .126 .114 .111 .118 .114 .114 .110 .119 .114 .106 .122 .116 .151 .143 Luxembourg .066 .095 .153 .160 .126 .172 .146 .124 .132 .144 .129 .131 .122 .141 Netherlands .135 .122 .164 .134 .119 .197 .123 .137 .163 .146 .127 .165 .152 .152 Portugal .050 .054 .043 .066 .052 .062 .042 .090 .120 .126 .117 .137 .116 .113 Spain .099 .101 .131 .133 .141 .132 .094 .096 .117 .125 .117 .121 .119 .132
15 hypothesis of the equality of means test which states that crisis cash holdings are larger than pre-crisis cash holdings. It is striking that the mean increased instead of decreased as was expected. However, this notion should be treated with caution as the result is not significant at any level.
Table 4. Equality of means test for the crisis
This table compares the means of cash ratios in the pre-crisis period and the crisis period, without taking into account any other variables.
Group Observations Mean Cash ratio
Pre-crisis 3,485 0.1350
Crisis 7,789 0.1370
Difference = mean(Pre-crisis) – mean(Crisis) Ha: diff < 0 Pr(T<t) = 0.2507
To look into the effect of multinationality, a test for equality of means is used again. The variance ratio test rejects the null hypothesis that the standard deviations are equal and thus a test for equality of means with unequal variances is used again. The results show that the mean for multinational firms is clearly lower in comparison to domestic firms. This is
confirmed by the acceptance of the alternative hypothesis that the mean cash ratio of domestic companies is larger compared to multinational firms, which is significant at a 0.01
significance level. This t- test thus provides preliminary evidence that the cashratios of multinational companies are smaller than the cash ratios of domestic companies.
Table 5. Equality of means for multinational firms
This table compares the means of cash ratios of domestic firms and multinational firms, without taking into account any other variables.
Group Observations Mean Cash ratio
Domestic firms 2,500 0.1466
Multinational firms 7,224 0.1327
16 As the report of the CPB (2014) in the introduction mentions higher cash holdings in the Netherlands compared to the core-EUR, this will now be looked into in more detail. The difference will be examined by performing equality of means tests between the countries of the Core-EU (Germany, France, Belgium, Italy) and the Netherlands. These tests are
performed to determine whether cash ratios in the Netherlands differ statistically significant from the other countries. First of all a variance ratio test rejected the null hypothesis that the standard deviations are equal for all countries, thus a test for equality of means with unequal variances is used. The outcome of these tests (see appendix, table 13-16) show that the cash holdings in the Netherlands are significantly smaller than cash holdings in Germany and significantly larger than cash holdings in Italy. There are no significant differences found between the cash holdings in Belgium, France and the Netherlands. It has to be noted here that no other variables are taken into account, so these tests only provide preliminary evidence.
Variables
In this section the different variables will be described, exact definitions and sources can be found in the appendix, table 17. All variables are denoted in Euro’s, except for GDP per capita. Moreover, all variables were examined for possible non normality and outliers. This resulted in the exclusion of several outliers and the usage of the logarithm of size.
The cash ratio will be the dependent variable, which is defined as cash divided by total assets, this is in line with Pinkowitz et al. (2016) and Bates et al. (2009) and is sourced from
Datastream.
The degree of multinationality consists of the amount of foreign sales divided by the total sales of a company. A multinational company is a company with more than 25 percent foreign sales. This definition is sourced from Worldscope by Pinkowitz et al. (2016) and is in line with Ferrnandes and Gonenç (2016) and Pinkowitz et al. (2012). When a company is a multinational company in one year, it will be a multinational for all years in the sample, which is in line with Pinkowitz et al., 2016.
17 variable is used that becomes 1 if a firm pays dividend if a firm does not the variable is 0. Leverage is measured by total debt to assets. It is expected that the substitutes are negatively related to cash holdings, as higher substitutes reduce the incentive to hold cash.
To measure firm size, the logarithm of total assets is used. The relationship between firm size and cash holdings is expected to be negative, as there are economies of scale with cash management and thus larger firms need relatively less cash. A logarithm is used as the distribution of firm size was non-normal.
To measure investments, R&D expenses to sales and capital expenditure to assets are used. When firms have no value for R&D expenses, then tis variable is set to 0, in line with Bates et al. (2009). The variable acquisition was not available in various databases. A positive
relationship is expected, as firms that invest much need money to do these investments. Firm characteristics that cover the extent to which firms are financially constrained are leverage and the industry’s cash flow volatility. A positive relationship is expected here, as more financially constrained firms face more difficulties when attracting external capital, therefore they need more cash to stay in business.
The market to book value of the assets serve as a proxy for the investment opportunities of a company. The better the investment opportunities, the more cash a firm is expected to hold, as adverse shocks and financial distress are more costly (Bates et al., 2009)
The controls for firms that raise capital are next. There are controls for equity and debt issues because capital raising firms tend to have more cash after they raised capital and their cash decreases again when they spend the capital they raised (Bates et al., 2009). Net debt issuance is measured by the long term debt due in one year plus total long term debt minus a lagged long term debt due in one year and a lagged total long term debt. Net equity issuance is measured by net proceeds from sale divided by the issue of common and preferred stock, which is available on Datastream.
18 criticize the WGI index, because the indicators are based on an average of zero. This means that if a value changes in one country, then values of other countries change as well.
Consequently, values of countries can change, while there is no actual change in the country. However, the WGI consist of much information and they give a general view of the
institutional quality of a country, which does make them useful in research.
In appendix, table 11 a correlation matrix for the WGI variables is provided, as the six indicators are summarized in one variable. The correlation values between the different WGI variables is varying from 0,55 to 0,96, which indicates that the WGI are subject to
multicollinearity. This is not surprising, as the different variables are highly correlated because they all measure a component of the quality of the institutions and governance in a country. Although Pinkowitz et al. (2016) do not take this into account, other research does. For example Globerman and Shapiro (2002) and Buchanan et al. (2011) do address this problem by using a principal component factor analysis. This approach is followed here as well and the results of this approach can be found in the appendix, table 12. This table shows that only the first factor can be retained, as its eigenvalue is 4.84, while the other factors have eigenvalues below 0.55. In table 12 the factor loadings are shown, from this table it can be concluded that political stability reports a high uniqueness value compared to the other variables with a value of 43.78 percent. This means that more than 43 percent of the variance in the variable is not explained by the factor.
The next country characteristics that are taken into account is the anti-self-dealing index (ASDI) from Djankov et al. (2008) which measures the extent to which minority shareholders are legally protected against expropriation by employees of the firm where they are
shareholders. A similar variable is the revised anti-director index (RADI), which was originally created by La Porta (1998) and revised by Djankov et al. (2008). This index is a measure of investor protection and the data can be found on the website of Andrei Shleifer1. Data is different for all countries, but does not change over time.
The measures of financial and economic development are discussed now. The development of the corporate bond market can be measured by bond market capitalization to GDP. It is expected that a more effective bond market will lead to easier access to external capital and thus the need for cash holdings will be smaller when bond market capitalization to GDP is
19 high. This data is sourced indirectly from the World bank and is available for all countries in the sample until 20112.
The second ratio is stock market trading divided by the stock market capitalization. This is a measure for the activity of the stock market and a proxy for stock market development. A better developed stock market is expected to lead to lower cash holdings, as this increases the availability of external capital. Bank credit is used as a proxy for development of the banking sector. A better developed banking sector is expected to lead to easier access to external capital and thus lower cash holdings. The last measure is GDP per capita in 2010 US dollars, this measure is included to make sure that no other variable corrects for changes in GDP per capita (Pinkowitz et al., 2016). These measures are sourced from the World Bank.
Method
First of all, both firm and country determinants will be taken into account to look into the effect of multinationality and to see whether the crisis has an influence on cash holdings. Net debt issuance and bond market had a large number of missing values, which reduced the available sample heavily. In order to increase the available data, these two variables were excluded from the final sample. A Hausman test was conducted to see if a fixed or random model was appropriate. The results (see appendix, table 9) implicated that a fixed model was appropriate to use. As there was a focus on the differences between countries, a country fixed effects regression with clustered standard errors is used. These standard errors are clustered by firm and take into account possible heteroscedasticity. To check for multicollinearity, a correlation matrix (see appendix, table 8) was made. There was one correlation above 0.5, this was the case between GDP and the WGI. However, as this correlation is not with the
independent variable and leaving them out would harm the model, it is chosen to
acknowledge that there may be multicollinearity in the model. Standard errors may increase because of this.
2 See:
20
Empirical Results
The outcome of the regression analysis is shown in table 7. The regression has an R-squared of 0.33, which means that 33 percent of the variance in the cash ratio is explained by the firm characteristics in the sample.
The results indicate that there is a positive relationship between being multinational and the cash ratio, as the positive coefficient of 0.0067 shows. This is contrary to the preliminary evidence provided by the equality of means test in table 5. However, as the probability value is 0.438, the result is not significant at any significance level. Therefore this results fails to confirm the first hypothesis, which stated that multinationality affects cash holdings.
Table 7. Country fixed effects regression with clustered robust standard errors
This table shows the output of the country fixed effects regression with clustered robust standard errors. Coefficients are shown next to the variables, standard errors are displayed in parentheses. Countries are compared to The Netherlands due to the dummy trap.
Variables Regression Output Countries (continued)
Multinational 0.0068 Austria -0.1163 (0.0087) (0.0224) Crisis 0.0011 Belgium -0.5201* (0.0037) (0.0788) Size -0.0068*** France -.1873 (0.0016) (0.040)*
Cash flow 0.0310 Germany 0.0085
(0.0280) (0.0228)
R&D 0.0001** Greece -0.1162*
(0.0000) (0.0354)
Capex -0.0011*** Ireland -0.7494*
(0.0003) (0.1008)
Equity Issuance 0.3360*** Italy -0.4421*
(0.0389) (0.720)
NWC -0.1720*** Luxembourg -0.2333*
(0.0185) (0.0487)
Leverage -0.3300*** Portugal -0.4241*
(0.0192) (0.0716)
Industry Volatility 0.0000* Slovakia -0.1388*
21 (0.0061) Total Observations 9,707 ASDI 1.7150*** R-squared 0.329 (0.2090) RADI -0.1180*** (0.0106) Bank Credit 0.0001 (0.0001) Turnover -0.0020 (0.0026) GDP 0.0000 (0.0000)
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
When looking at the other probability values, it stands out that the dummy variable to take into account the crisis is insignificant. This is in line with the preliminary evidence found in the equality of means test is table 4, where it was found that pre-crisis cash ratios were not statistically significantly different from crisis cash ratios. The other variables are statistically significant, except for cash flow, MTB, Dividend, WGI, turnover, bank credit and GDP.
The results show that size has a negative and significant coefficient. This indicates that there are economies of scale with cash holdings and the results are in line with both the transaction motive and the precautionary motive.
The coefficient for the investment characteristic R&D is positively related to the cash ratio. For R&D the coefficient is relatively small considering this variable consist of R&D
expenditures divided by sales, however the coefficient is significant. Capex, a variable measuring capital expenditure, is the other investment characteristic taken into account. Contrary to R&D, capex is negatively related to the cash ratio, which is contrary to what Opler et al. (1999) find. Both the precautionary motive and the agency motive led to the expectation of a positive relationship between investment and the cash ratio, which is only supported by the coefficient of R&D.
22 substitutes of cash holdings, there is less need to hold cash, which is in line with the
precautionary motive.
Equity issuance is positively related to the cash ratio. This suggests that when firms have raised capital, they hold more cash, which is in line with the precautionary motive. The industry volatility is negatively related to the cash ratio. However, the corresponding
coefficient is very low. The negative relationship is contrary to the expectation that firms with higher industry cash flow risk were expected to have higher cash ratios to compensate for the fact that they face greater difficulties in attracting external capital.
The ASDI coefficient is positive and significant. Better minority shareholder protection was expected to lead to a lower cash ratio, however the results indicate that this is the other way around. The coefficient of RADI is negatively related to the cash ratio and significant. This result implies that more investor protection leads to lower degrees of the cash ratio, which is in line with the agency motive.
The coefficients on the country fixed effects show the intercepts of firms in the countries in our sample relative to firms in the Netherlands, as fixed effects for the Netherlands are not included due to the dummy trap. It stands out that all countries except Germany have smaller coefficients than The Netherlands. The coefficients are significant for all countries except for Germany, Spain and Austria. These results correspond to the equality to the equality of means tests for Germany and Italy that were executed earlier (see appendix, table 13-16). For
Belgium and France, the equality of means tests didn’t show a significant difference in means between the Netherlands and France and Belgium, but the regression results which control for time-varying firm characteristics and time-varying country characteristics show that Dutch firms have significantly higher average cash ratios than firms from these countries, as well as compared to firms from Greece, Ireland, Portugal, Luxembourg and Slovakia. This suggests that non-time varying factors in the Netherlands result in significantly higher cash ratios compared to eight of the other countries in the sample. Potential non-time varying factors that could affect cash holdings of Dutch firms could be features of the tax system or the
23
Conclusion and limitations
This research looked into corporate cash holdings in the European Union. Preliminary evidence seemed to point at a negative relationship between multinationality and cash holdings. However, in the regression analysis the effect of multinationality on cash holdings in the European Union was not found to be significant. This is not striking as mixed results were provided in earlier research. With regard to the effect of the crisis on cash holdings, the overall development of the cash ratio over the sample, as shown in figure 1, shows that the cash ratio was quite stable in the years prior to the crisis. When taking into account the effect of the crisis, it stands out that the start of the crisis in 2007 decreased cash holdings of firms all over Europe. This was true for every country in the sample. However, this effect was only visible for one year, in 2009 all countries except for Luxembourg reported growing cash holdings compared to the prior year. Even more striking, in recent years (from 2012 to 2015) the cash holdings even started to grow to levels that were even higher than before the crisis. When looking at the differences between European countries, it stands out that there are significant differences among countries with regard to cash holdings. The regression
compared all countries in the sample to the Netherlands, findings include that Portugal, Italy, France, Greece, Slovakia, Ireland, Luxembourg and Belgium all have negative significant coefficients. When looking into the difference that between the core-EU and The Netherlands mentioned in the introduction, preliminary evidence provided no statistical proof that The Netherlands has higher cash holdings compared to countries in the core-EU, which included Germany, Belgium, France and Italy. The Netherlands only reports statistically significant higher cash holdings compared to Italy. When comparing to Germany, the Netherlands
reports significantly smaller cash holdings, while the differences with Belgium and France are insignificant. However, when taking into account the determinants of cash holdings, it is striking to see that the Netherlands has significantly higher cash ratios than eight out of eleven countries in the sample. This means that being just the fact that a firm is Dutch increases the cash holdings of a company.
24 see that within Europe there are differences between countries, that cannot be explained by current literature on cash holdings.
One of the limitations of this research is that the variable acquisition to assets was not available. This variable is seen as one of the determinants of cash holdings and was used by many papers such as Pinkowitz et al. (2016), Fernandes and Gonenç (2016) and Bates et al. (2009). Moreover, the variables net debt issuance and bond market were not used although they were available. The usage of these variables would decrease the sample substantially, which is why they were dropped.
Directions for future research could include a focus on what characteristics of certain firms or countries determine cash holdings. As shown in this paper, many factors are already known, but the regression results presented show significant cross-country differences when
controlling for time-varying firm and country characteristics. Exploring the factors driving these differences, such as other firm characteristics or other (non-time varying) country characteristics, is a question for future research.
25
References
Almeida, H., Campello, M., Laranjeira, B., & Weisbenner, S. (2009). Corporate debt maturity and the real effects of the 2007 credit crisis (No. w14990). National Bureau of Economic Research.
Bates, T. W., Kahle, K. M., & Stulz, R. M. (2009). Why do US firms hold so much more cash than they used to?. The journal of finance, 64(5), 1985-2021.
Campello, M., Graham, J. R., & Harvey, C. R. (2010). The real effects of financial
constraints: Evidence from a financial crisis. Journal of financial Economics, 97(3), 470-487.
Chen, L., & Zhao, X. (2006). On the relation between the market-to-book ratio, growth opportunity, and leverage ratio. Finance Research Letters, 3(4), 253-266.
Denis, D. J., & Sibilkov, V. (2009). Financial constraints, investment, and the value of cash holdings. Review of financial studies, hhp031.
Dittmar, A., Mahrt-Smith, J., & Servaes, H. (2003). International corporate governance and corporate cash holdings. Journal of Financial and Quantitative analysis, 38(01), 111-133.
Fernandes, N., & Gonenc, H. (2016). Multinationals and cash holdings. Journal of Corporate Finance, 39, 139-154.
Ferreira, M. A., & Vilela, A. S. (2004). Why do firms hold cash? Evidence from EMU countries. European Financial Management, 10(2), 295-319.
Foley, C. F., Hartzell, J. C., Titman, S., & Twite, G. (2007). Why do firms hold so much cash? A tax-based explanation. Journal of Financial Economics, 86(3), 579-607.
Fresard, L. (2010). Financial strength and product market behavior: The real effects of corporate cash holdings. The Journal of finance, 65(3), 1097-1122.
26 Harford, J. (1999). Corporate cash reserves and acquisitions. The Journal of Finance, 54(6), 1969-1997.
Haushalter, D., Klasa, S., & Maxwell, W. F. (2007). The influence of product market dynamics on a firm's cash holdings and hedging behavior. Journal of Financial Economics, 84(3), 797-825.
Jansen, C., & Ligthart, M. (2014). Spaaroverschot niet-financiële bedrijven: ontwikkeling, oorzaken en gevolgen. Retrieved February 17, 2017, from
https://www.cpb.nl/publicatie/spaaroverschot-niet-financi%C3%ABle-bedrijven-ontwikkeling-oorzaken-en-gevolgen
Keynes, J.M. (1936). The General Theory of Employment. Interest and Money. Hartcourt Brace, London.
Porta, R. L., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. W. (1998). Law and finance. Journal of political economy, 106(6), 1113-1155.
Love, I. (2001). Financial development and financing constraints: international evidence from the structural investment model (Vol. 2694). World Bank, Development Research Group, Finance.
Moodys: US corporate cash pile, led by tech sector, to grow to $1.77 trillion by end of 2016. (2016, November 03). Retrieved April 12, 2017, from
https://www.moodys.com/research/Moodys-US-corporate-cash-pile-led-by-tech-sector-to--PR_357576
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.
Oman, C., & Arndt, C. (2006). Governance Indicators for Development (No. 33). OECD Publishing.
27 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. The Journal of Finance, 61(6), 2725-2751.
Pinkowitz, L., Stulz, R. M., & Williamson, R. (2016). Do US firms hold more cash than foreign firms do?. Review of Financial Studies, 29(2), 309-348.
Pinkowitz, L., Stulz, R. M., & Williamson, R. (2012). Multinationals and the high cash holdings puzzle (No. w18120). National Bureau of Economic Research.
Still slipping the net. (2015, October 08). Retrieved April 06, 2017, from
Appendix
Table 8. Correlation matrix of cash holding determinants
This table consists of the correlations between all variables used in the regression in table 7. Correlations above 0.5 could implicate multicollinearity. This is only the case with GDP and the WGI however, to make sure the model is not harmed both variables are included in the model.
Cash ratio
Multinational Crisis Size Cash
flow
R&D Capex Equity
29 Table 8. Correlation matrix firm characteristics (continued)
(Continued) Industry Volatility MTB Dividend WGI ASDI RADI Bank Credit Turnover GDP
Table 9. Hausman test
The outcome of the Hausman test shows that it is appropriate to use a fixed model for the regression. The null hypothesis of random effects is rejected at a 0.01 significance level.
chi2(6) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 231,84
Prob>chi2 = 0.0000
Table 10 Correlation Matrix WGI
This table shows the correlations between the different variables that form the WGI. Voice and Accountability Rule of Law Regulatory Quality Political Stability Government Effectiveness Control of Corruption Voice and Accountability 1 Rule of Law 0.7287 1 Regulatory Quality 0.7460 0.9054 1 Political Stability 0.5541 0.6422 0.6977 1 Government Effectiveness 0.7166 0.8868 0.8056 0.5654 1 Control of Corruption 0.7471 0.9598 0.9188 0.6465 0.897 1
Table 11. Principal component factor analysis results
Factor Eigenvalue Difference Proportion Cumulative
31 Table 12. Factor loadings (pattern matrix) and unique variances
Variable Factor1 Uniqueness
Voice and Accountability 0.8326 0.3068
Rule of Law 0.9591 0.0801
Regulatory Quality 0.9470 0.1032
Political Stability 0.7498 0.4378
Government Effectiveness 0.9120 0.1683
Control of Corruption 0.9676 0.0638
Table 13. Equality of means test Germany-Netherlands
This table compares the means of cash ratios in The Netherlands and Germany, without taking into account any other variables.
Group Observations Mean
Germany 2,612 0.1646
The Netherlands 239 0.1427
Diff = mean(German firms) – mean (Dutch firms) Ha: diff > 0 Pr(T>t) = 0.0112
Table 14.Equality of means test Belgium-Netherlands
This table compares the means of cash ratios in The Netherlands and Belgium, without taking into account any other variables.
Group Observations Mean
Belgium 740 0.1452
The Netherlands 239 0.1427
32 Table 15. Equality of means test France-Netherlands
This table compares the means of cash ratios in The Netherlands and France, without taking into account any other variables.
Group Observations Mean
France 2995 0.1459
The Netherlands 239 0.1427
Diff = mean(French firms) – mean (Dutch firms) Ha: diff != 0 Pr(|T| > |t|) = 0.7343
Table 16. Equality of means test Italy Netherlands
This table compares the means of cash ratios in The Netherlands and Italy, without taking into account any other variables.
Group Observations Mean
Italy 383 0.1175
The Netherlands 239 0.1427
Diff = mean(Italy firms) – mean (Dutch firms) Ha: diff < 0 Pr(T < t) = 0.0074
Table 17. List of variables
ADRI Revised anti-director rights index, from:
https://scholar.harvard.edu/shleifer/publications?page=3 ASDI Anti-self-dealing index, from:
https://scholar.harvard.edu/shleifer/publications?page=3
Bank Credit Bank credit to GDP, sourced from the World bank Bond Market Private Bond Market Capitalization to GDP, from:
33 Capex Capital expenditures divided by assets, from: Datastream
Cash Cash divided by assets, from: Datastream Cash flow Cash flow divided by assets, from: Datastream
Debt Issuance Net debt issuance divided by assets. Net debt issuance is computed by: (long term debt due in one year + total long term debt) – (lagged long term debt due in one year + lagged total long term debt). From: Datastream
Dividend Dummy variable, turns into 1if a firm pays dividend, from: Datastream Equity Issuance Net equity issuance divided by assets. Net equity issuance consists of
net proceeds from sale and issue of common & preferred equity stock. From: Datastream
GDP GDP per capita in 2010 US dollars, from: World bank
Industry volatility Industry mean of firm standard deviation of the cash flow of the prior ten years, with a minimum of three years data availability to compute firm volatility, from: Datastream
Leverage Total debt to assets. Total debt is computed by short term debt + long term debt, from: Datastream
MTB Market to book value of total assets. Computed by: ((total assets - total equity) + (common shares outstanding * share price)) / total assets. From: Datastream.
MNC Dummy variable which turns into one if a firm is a multinational, firms are multinational when 25% or more of their sales are from abroad, from: Datastream.
NWC Net working capital to total assets, computed by: (working capital – cash and short term investments) / total assets. From: Datastream. R&D Research and Development to total sales, from Datastream Size Logarithm of total assets, from: Datastream
Turnover Stock market turnover, from: World bank
