Cash Holdings of European Firms. Abstract

Transcription

1 FINANCE MASTER THESIS VOL. I, NO. I OCTOBER 2011 Cash Holdings of European Firms. Jusmir Jasarevic Tilburg School of Economics and Management Tilburg University Msc. Finance CFA-Track ANR: U Supervisor Vasso Ioannidou CentER, Tilburg University Abstract In this paper I examine the evolution of cash holdings for European firms for the period from 1989 to I find a positive and significant increase of the cash ratio over time. The academic literature distinguishes three motives for firms to hold cash: the transaction motive, the precautionary motive, and the agency motive. I find evidence consistent with the existence of the transaction motive as larger firms hold more cash and firms with higher capital expenditure have lower cash holdings. The results show no consistent evidence that the precautionary motive contributes to the determination of cash holdings. Furthermore, I find evidence supporting the agency motive in countries where corporate governance is weaker. In particular, firms incorporated in countries with weak corporate governance hold more cash. * I would like to thank my supervisor Vasso Ioannidou for guidance during the writing of this thesis and helpful comments. I also want to thank my parents for giving me the opportunity to study even after they lost everything in the war their mission in life was to give me the best chances possible. A word of thanks goes to Marnix Koster and Mina Shahabi for correcting the first view versions of this thesis. The views expressed do not necessarily reflect those of the people mentioned above or the University of Tilburg. All remaining errors are all mine. 1

2 2 Finance Master Thesis Table of Contents 1. Introduction Literature review Data and information environment Empirical analysis Results Conclusion References Appendix... 26

3 Cash Holdings of European Firms 3 1. Introduction Firms cash holdings remain an interesting topic for the financial media. An example is an article retrieved from Reuters.com which describes the cash stock piles in U.S. firms after the financial crisis: U.S. companies are hoarding almost $ 1 trillion in cash unlikely to spend due to continuing uncertainty about strength of the economy. 1 This hoarding of cash is also done by European firms, they hold about $ 700 billion in cash as they head to Holding cash is costly for firms as cash on its own does not earn the required rate of return. However, these cash holdings could lower distress cost and thus offset the cost of holding cash. Cash holdings can also be a determinant of agency problems. As Jensen (1986) puts it, firms with agency problems are expected to accumulate cash when there are no good investment opportunities and its management does not want to return it to their shareholders through dividends or share repurchases. Because of large technical improvements in the credit markets, a decrease in cash holdings can be expected. The availability of more and better hedging opportunities could realize this reduction. However, Bates, Kahle and Stulz (2009), henceforth BKS, found that the cash ratio more than doubled. During their sample period of 1980 to 2006 for U.S. firms, the ratio went from 10.5% to 23.2%. Therefore, they found it important to investigate whether this increase is due to agency problems, represent an anomaly that challenges existing theories of the determinants of corporate cash holdings, or results from changes in firm characteristics and their business environment. The paper by BKS limits their sample to the cash holdings of U.S. firms. As shown by the above articles, European firms hold almost the same amount of cash as U.S. firms. Even though they hold almost the same amount of cash, no prior research has been done on the determinants of cash holdings of European firms. This paper investigates the cash holdings of European firms since It also investigates whether known determinants of cash holdings can explain the evolution of cash holdings. Where possible this paper will compare its results to the findings of BKS for the U.S. Furthermore, this paper will extend their sample period until Therefore, the financial crisis will be included along with its possible effects. Sample differences in Europe are also taken into account. These differences are in a governments economic model (Anglo-Saxon vs. Continental Capitalism) and 1 Karen Brettel, U.S. companies hoarding almost $ 1 trillion cash: Moody`s, Reuters.com, October 26, Richard Weis and Francois de Beaupuy, European Companies With $ 700 Billion Cash Ease Deal Dearth, Bloomberg.com, November 30, 2010.

4 4 Finance Master Thesis governments financial status (AAA-rated and those that are not AAA-rated). I will use financial statement information that is provided by all firms. This gives reliable financial figures because their end of year accounting publications are checked by independent accounting firms. This paper will proceed as follows. Section 2 gives an theoretical analyses of the determinants of cash holdings. Section 3 describes the financial and market data. Section 4 describes the empirical test. Section 5 presents the results and section 6 concludes. 2. Literature review The optimal capital structure and the way it is chosen by companies, is a fundamental question in financial economics (Lemmon, Roberts & Zender, 2008). A piece of this puzzle is the cash holdings of firms. The importance of this piece is particularly shown in the paper by BKS, as they use cash to generate a different measure for leverage than used in the financial literature. They use net debt ratio, which is commonly used by practitioners, instead of debt ratio, which is used in the financial literature. The difference is that, when using net debt, cash is subtracted from debt before dividing it by total assets or equity. The use of net debt for leverage leads them to conclude: dramatically different conclusions about both the current level of leverage in U.S. firms and the evolution of leverage over the last 25years. In a Modigliani-Miller world, firms would be indifferent if they would hold cash or borrow money in case of a shortage. In this case the cost of capital is independent of the debtequity ratio (Modigliani & Miller, 1958). However, in the world we live in, it is costly for firms to have a shortage of liquidity. Therefore a firms cash level should be set so that the marginal benefit, avoiding the cost of a shortage, equals the marginal costs of holding cash. The cost of holding cash is called the liquidity premium and is the difference between the firms rate of return and the lower rate of return on its cash holding. The financial theory provides three classic motives for firms to hold cash. Keynes (1936), describes the first two motives as the transaction motive and the precautionary motive. The third motive is added by Jensen and Meckling (1976) and is the agency motive. The Transaction motive. This motive refers to the transaction cost a firm encounters when converting noncash financial assets into cash. Firms need cash for operations and for financing its operations. When a firm is short of liquidity it has to raise funds through either

5 Cash Holdings of European Firms 5 accessing the capital market, liquidate assets, cut dividends, lower investments or renegotiate its current financial contracts. Indifferent of the options, all are costly for the firm. As Pinkowitz, Stulz and Williamson (2003; pp. 3) state: it is cheaper to hold cash than to send somebody to the bank. Opler, Pinkowitz, Stulz and Williamson (1999), henceforth OPSW, show that a firm has an optimal amount of cash when marginal cost of holding cash equals the marginal cost of a cash shortage. This is graphically shown in figure 1, this figure is from their paper. Figure 1: Optimal holdings of liquid assets(opler, Pinkowitz, Stulz & Williamson (1999)). The intersection of the marginal cost of liquid assets and the marginal cost of liquid assets shortage is the optimal amount of liquid assets. Marginal cost of holding liquid assets is assumed to be non-decreasing, this represents the horizontal curve. The decreasing curve represent the marginal cost of liquid asset shortage. The decrease in marginal cost of a shortage in liquid assets is due to the fact that a larger shortage leads to higher costs to a firm. This is because the firm first uses the least costly measure to gather funds, but in case that is insufficient, it will need to apply more costly measures. The costs of holding cash are assumed to be constant and therefore represent a horizontal curve. Mulligan (1997) provides evidence supporting economies of scale with regard to the transaction motive. This means that large firms hold relatively less cash.

6 6 Finance Master Thesis The Precautionary motive. The precautionary demand for cash theory goes back to 1888, when it was first mentioned by Edgeworth (Baltensperger, 1974). This theory states that firms hold cash because of uncertainties in firm cash flows. When access to capital markets is costly for firms, they hold cash to better withstand these shocks in cash flows. The costs for accessing the capital markets comes from information asymmetries. Information asymmetries makes outsiders discount firms securities to protect themselves. If discounts are large enough, it can be more interesting and profitable for a firm to reduce investments instead of selling securities (Myers & Majluf, 1984). The precautionary motive is therefore based on the information asymmetry and a firm s ability to raise funds. This brings up the pecking order theory which states that firms first preference is the use of internal funds to finance investment, their second preference is the issuance of debt, and their least preferred option is to issue equity. OPSW find evidence that firms with poor access to external capital markets and with riskier cash flows hold more cash. Another suggestion by the precautionary motive is that firms with better investment opportunities hold more cash as an adverse shock would be more costly for them. OPSW find support for this prediction by using market-to-book and R&D expenses as a proxy for investment opportunity. The precautionary motive is modeled in the paper by Han and Qiu (2007). They theoretically show and empirically prove, that cash holdings of constrained firms increase with an increase of firms cash flow volatility. The Agency motive. Agency problems occur when a firms management (agents) and its shareholders (principals) have conflicting interests. The presence of agency problems can be find by looking at cash holdings of firms. The free cash flow theory by Jensen (1986) states that entrenched managers accumulate more cash instead of paying out to shareholders even when there are no good investments opportunities (see also Stulz, 1990). Entrenched managers use this excess cash to pursue their own objectives at the cost of shareholders. Jensen (1986) indicates managers power and compensation as an incentive for them to hold excessive cash. OPSW also points out that holding cash makes managers more risk averse as it can avoid market discipline. Market discipline occurs when a company has to access capital markets to obtain new capital. By holdings cash, the firms avoid monitoring of capital markets with respect to the firms

7 Cash Holdings of European Firms 7 operations and investments. Managerial perks are another reason for holding excess cash. Excess cash leads firms to engage in, as famous and successful investor Peter Lynch calls it, diworsification 3. It means that diversification policies mostly do more harm than good to a firm. Evidence supporting this can be found in Blanchard, Lopezde-Silanes and Shleifer (1994) and Harford (1999). Harford finds that cash-rich firms are more likely to make acquisitions and that these acquisitions are value decreasing. He also finds that the period after the cash-rich acquisition is followed by an abnormal decline in operating performance. Agency problems are costly for shareholders as shown above and therefore it is important to look at corporate governance. Corporate governance is a proxy for agency problems as it presents the shareholders rights. If shareholders rights are less protected by law, it gives them less power to discipline their managers. This gives more freedom for entrenched management and therefore creates agency problems. Evidence for this can be found in the paper by Dittmar, Mahrt-Smith and Servaes (2003). They find that firms hold up to twice as much cash in countries where shareholders rights are not well protected, compared to firms in countries with good shareholder protection. Kalcheva and Lins (2007) find that cash is less worth in countries where shareholder rights are less protected and firms hold more excess cash. Harford (1999) finds that cash-rich bidder firms destroy up to seven cent of every excess dollar they hold. In line with these findings Dittmar and Mahrt-Smith (2007) also find that corporate governance has an substantial impact on the value of cash in a firm. They find that in poorly governed firms excess cash is discounted up to $ 0.42 cent per dollar hold. Their results furthermore show that firms with poor governance spend excess cash quickly (see also Harford, Mansi & Maxwell, 2008). All these papers provide evidence that agency problems are costly for shareholders. Investors discount excess cash in firms with poor governance, this is because they have less power to control the managers to distribute the cash back to shareholders. Also their power to discipline managers for bad investment and managerial perks is lower, which makes them discount the cash as it has less value for them. Adding to the discounting of excess cash is that firms with poor governance tend to spend excess cash quickly, this gives shareholders less time to evaluate investments and less time to force managers to distribute cash back to shareholders. Frésard and Salva (2010) find the opposite to be true regarding to firms with good corporate governance. The value 3 Peter Lynch, One Up on Wall Street, SIMON & SCHUSTER PAPERBACKS, 2000.

8 8 Finance Master Thesis of excess cash for foreign firms that are listed in the U.S. are substantially larger than for their domestic peers. Better shareholders rights and disclosure requirements constrain management on inefficient allocation of corporate resources. The value of excess cash is determined by the expectation on where it is used for. Better corporate governance in the U.S. gives them more power to control this expectation and therefore they give it a higher value. 3. Data and information environment This paper uses data from Compustat Global, which is available through Wharton Research Data Services (WRDS). The data provides authoritative financial and market data of listed companies, U.S. and Canada excluded, in more than 80 countries. The data has only valuable information from Therefore my sample covers the period between 1989 and This paper sample period ranges till 2010, so it includes the financial crisis. The study on U.S. firms by BKS does not include the crisis and its possible effects as their sample period stops in For each firm, I have financial information which can be categorized in Company Financials and Financial Markets. Company Financials provide information on the income statement, balance sheet, flow of funds and other supplement data items. Financial Markets give information on stock market related items. However, this data is missing for a lot of the firms in the sample. I therefore use Datastream Worldscope to obtain the data on market capitalization and paid dividends. The sample used in this paper includes all firms that are incorporated in one of the countries that were member of the European Union in In 1995 the European Union had a total of 15 membership states. Additional to these countries I also add Switzerland to my sample. However, Switzerland is not an member of the European Union, it is located in the central of Europe and all neighbors are European Members. All these countries are classified as advanced economies by the International Monetary Fund 5. Despite this classification, there is an in sample variance between the financial health status of countries, for example Netherlands (AAA) compared to Greece (CC). This sample also includes two different types of economic models, Anglo-Saxon (United Kingdom and Ireland) and Continental Capitalism 4 Member states: Belgium, Germany, France, Italy, Luxembourg, Netherlands, Denmark, Ireland, United Kingdom, Greece, Spain, Portugal, Austria, Finland and Sweden. Source: page 67.

9 Cash Holdings of European Firms 9 (continental Europe). Other restrictions for the sample are that firms have positive assets and positive sales. Financial firms are excluded (SIC codes ) as so are utilities (SIC codes ). This due to the fact that financial firms can carry cash to meet capital requirements and the cash holdings of utilities can be under regulatory supervision. Table 1 gives the variables used in this study along with their definitions. Insert Table 1 near here The sample yields a panel of 71,619 observations for 6,812 unique firms. This includes surviving and non-surviving firms. Table 2 provides descriptive statistics on the variables used in this study. Insert Table 2 near here Outliers are winsorized for several of the variables to delete extreme outliers that produce noise in the regressions. Market-to-book is winsorized at 1% level on the top tail. Leverage is winsorized so that it is between zero and one. Capital expenditure/assets is winsorized at the 1% level. Cash flow/assets and NWC/assets are winsorized at the 1% level on the bottom tails. In Table 3 the average and median cash and leverage ratio is shown for our sample. The second column reports the number of unique firms for each year. The number of unique firms drops by 27.5% from 2006 to 2010 which can be a direct effect of the financial crisis. Cash ratio is a measure of cash and marketable securities divided by total assets. The aggregate cash ratio is the sum of cash divided by the sum of assets for all sample firms and is presented in the third column of Table 3. This aggregate cash ratio starts at 13.7% in 1989 and drops to 10.2% in 2010 with an average of 11% during the sample period and a peak of 16.2% in 1997 and a low of 6.6% in Column four of Table 3 reports the average cash ratio by year. The cash ratio increases from 11.4% in 1989 to 15.5% in 2010 with a peak of 16.8% in The median cash ratio is reported in column 5 and has the same rising trend as conveyed in column 4. The average and median cash ratio show a decrease in 2007 and 2008, which where the years of the crisis. After these years the cash ratio started to rise again in line with the trend until I regress the cash ratio on a constant and time to asses if there is a statistically significant trend. I first regress pre-crisis (until 2006) and show the U.S. results in brackets. The yearly average increase of the average cash ratio is 0.36% (0.46%) and has a ᴘ- value below 0.01 (<0.01). The regressions R² is 90% (89%). I find for the median an yearly average increase of 0.09% (0.28%) with a ᴘ-value below 0.01 (<0.01) and R² of 43% (70%). These results show a positive time trend in cash holdings over the sample period. The evidence is stronger for the US sample as the results have a larger positive and significant

10 10 Finance Master Thesis coefficient for the average- as well as the median cash ratio. When regressing with the full sample the coefficient of the yearly average increase of the cash ratio is 0.27% with an R² is 79% and ᴘ-value below This is 0.09% lower than the pre-crisis coefficient. The median cash ratio does not change post crisis. Insert Table 3 near here The average debt for the sample firm is reported in column 6 of Table 3. The measurement for debt is long-term debt plus current liabilities, divided by book assets. Leverage ratio stays stable around 22% and the median of the leverage ratio is reported in column 7. When using net leverage, which subtracts cash from debt, BKS comes to dramatically different perspective regarding time trend in leverage for U.S. firms. They find that the net debt ratio is 16.4% in 1980 and falls during the next years to become negative in the last 3 years of their sample. When looking at the European sample, I do not find such trend. I see that between 1992 till 2000 there is a falling trend but this trend is reversed in the next 3 years. After those 3 years there is again a downward trend from 2004 till After this decrease the net leverage is again reversed in Overall there is no clear trend in the European sample and differs therefore from the U.S. sample. In a regression on a constant and time the coefficient for the average net leverage is -0.32% (-0.60%) and a ᴘ-value below 0.01 (<0.01) pre-crisis. Post-crisis the coefficient is -0.21% and a ᴘ-value below One possible explanation for the two reversions of net leverage is that it occurs right after a crisis. In 2000 this was the Dot-Com Bubble and in 2007/8 this is the Financial crisis. The last column reports the median net leverage and experiences a downward trend for the European sample as well as the U.S. sample. Next, I divide the sample into AAA-rated and lower than AAA-rated countries. After the financial crisis some of the European governments came in to financial distress and their rating was adjusted downwards. By dividing the sample it gives the possibility to analyze whether a countries credit rating effects the cash holdings of firms. The firms with an AAArating are: Germany, France, Luxembourg, Netherlands, Denmark, United Kingdom, Austria, Finland and Sweden. The firms with a rating lower than AAA: Belgium, Italy, Ireland,

11 Cash Holdings of European Firms 11 Greece, Spain, Portugal. 6 Additional information that is known to all in finance, but crucial to emphasize, is that the U.S. government is AAA-rated. 7 Insert Table 4 near here Table 4 shows the same variables as Table 3, which is presented above. Additional to the results of Table 3, which are repeated in the first line of each column, I add the results from the separated sample which are stated in the second and third line. The second line of every column represent the results for only AAA-rated firms. The results for countries that do not have a AAA-rating can be found in the third line of the columns. Results from the average and median cash ratio show that the reported increase from the whole European sample purely results from firms in AAA-rated countries. Firms in countries where governments do not have an AAA-rating report a decrease, which is most clear for the median cash ratio. When running regression on the above results, on a constant and time, I find only statistically significant results for the AAA-rated sample. It results in a yearly average increase of the average cash ratio by 0.43% pre-crisis, and 0.32% post-crisis. Both regressions have a ᴘ-value below This shows that the AAA-rated sample coefficient is close to the one of the U.S. which is a 0.46% yearly average increase in cash ratio. The coefficient of the median gives an yearly average increase of 0.14% pre-crisis as well as postcrisis. Even though it is an increase from the full sample, 0.05%, it is still widely different from the U.S. sample (0.28%). The average net leverage coefficient reports a yearly average decrease of -0.42% pre-crisis and -0.32% post-crisis. This is an decreases in both coefficient compared to the full sample and brings the pre-crisis coefficient close to -0.60% of the U.S. sample in BKS. When regressing the sample that does not have an AAA-rating, it gives a positive and statistically significant coefficient for average net leverage, post crisis of 0.27%. Observing results from Table 4 and the regression above, one sees that firms incorporated in countries where governments have an AAA credit rating are holding more cash. One reason could be that if the financial status of an government is healthy, the country s economy is also healthy and therefore firms are performing better. Better performance leads to better results and therefore the ability to accumulate more cash. The above results illustrate a secular increase in the average cash ratio. Next I assess whether this increase is related to firm size. I do this by dividing the sample firms into 5 6 Standard & Poor s, RatingsDirect on the Global Credit Portal, August 5, 2011 Moody s Investor Service, Sovereign and Supranational Issuer Rating Summary, August 5, 2011 Fitch Rating, Sovereign Rating History, 12 September, S&P downgraded US Treasury Bonds to AA+, however the other two major rating agencies did not.

12 12 Finance Master Thesis quintiles each year based on size. The size measure is their book value of assets at the end of the prior year. Figure 2 illustrates the results graphically. 0.3 Q1: Smallest firm size quintile Q2 Q3 Q4 Q5: Largest firm size quintile 0.25 Cash ratio Figure 2: Average cash ratio divided in 5 firm size quintiles. The sample includes all Compustat firm-year observations from 1989 to 2010 with positive values for the book value of total assets and sales revenues for firms incorporated in the European Union (1995) and Switserland. Financial firms (SIC code ) and utilities (SIC code ) are excluded from the sample, yielding a panel of 71,643 for 6,812 firms. Cash ratio is a measure of cash and marketable securities divided by total asset. Firms are sorted into 5 quintiles based on their size. The size measure is the book value of assets at the end of the prior year. The first quintile (Q1) represents the smallest firms in the sample and the fifth quintile (Q5) the largest firms in the sample. The figure shows that smaller firms have the largest increase in cash ratio. The two quintiles for large firms, Q5 and Q4, show a slight decrease in the cash ratio. However, it is remarkable to see that the cash ratio of smaller firms is lower than those of big firms at the start of the sample period. BKS find an increase in the cash ratio for all their firm size quintiles. As expected the increase in cash holding is driven by small firms. In Table 5, I turn to the role of dividends and net negative income on the cash ratio. Columns 2 and 3 represent the average cash ratio for dividend and nondividend payers in a time series where column 4 and 5 show a time series of firms with negative net income and other firms. The paper by Fama and French (2001) provides evidence that dividends are disappearing and BKS finds that this is related to the increase in cash holdings. Firms are less likely to pay dividends which can be put in relation to the precautionary motive for firms to hold cash. But it can also be seen in respect to the agency motive, as entrenched management with poor growth opportunities will accumulate cash instead of distributing it to shareholders. Column 2 shows that the average cash ratio for dividend payers stayed around the same ratio Year

13 Cash Holdings of European Firms 13 throughout the sample. Nondividend payers however almost doubled (68%) their cash ratio pre-crisis. This is also observed for US firms in BKS which report an 113% higher cash ratio in 2006 compared to Insert Table 5 near here The cash ratio of firms with negative net income is 153% higher in 2006 than in 1989 and for nonnegative net income this is 43%. This shows that firms that experience a negative net income hold almost four times as much cash than those with a positive income. Firms that experience negative net income are considered to be more financially constrained. This can explain their increase in cash holdings from the precautionary motive. The cash ratio of firms with a negative income almost triples for BKS which is also the case for European firms. I next turn to the composition of the firms on the market. Technology firms are expected to hold more cash than old-economy manufacturing. High-tech firms invest more in R&D and are in a fast-paced industry where missing the newest development can be destructive. That is why it is expected that these firms hold more precautionary cash, as being short of it brings them high distress costs. Fama and French (2004) find a dramatically expansion of listing in the 1980s and 1990s on the U.S. stock exchange. It therefore could be that the above observed increase in cash holdings is the result of an increase on the proportion of high-tech firms in my sample. Old-economy manufacturing firms are defined as firms with SIC codes and technology firms by the definition used in Loughran and Ritter (2004). In Table 6, the proportion of high-tech firms compared to the full sample and the sample with high-tech and old-economy manufacturing firms is shown. Insert Table 6 near here The results show that the proportion of high-tech firms tripled over the sample period. In 1989 the proportion of high-tech firms was 7% and by 2010 this increased to 21%. The same increase holds when you look at the sub sample of only high-tech and old-economy manufacturing firms. I have now shown that the proportion of high-tech firms, which are expected to hold more cash, increased dramatically during the sample period. The next step is to look at the evolution of average cash ratio for high-tech and old-economy manufacturing firms which is shown in Figure 3.

14 14 Finance Master Thesis 0.33 "old-economy" manufacturing firms high-tech firms 0.28 Cash ratio Year Figure 3: Average cash ratio of high-tech and old-economy manufacturing firms. The sample includes all Compustat firm-year observations from 1989 to 2010 with positive values for the book value of total assets and sales revenues for firms incorporated in the European Union (1995) and Switserland. Financial firms (SIC code ) and utilities (SIC code ) are excluded from the sample. Additional to this the sample is divided into old-economy manufacturing firms with SIC-codes yielding a panel of 31,582 for 2,760 firms and high-tech firms yielding a panel of 13,127 for 1,401 firms. Cash ratio is a measure of cash and marketable securities divided by total asset. Figure 3 shows that during all sample years, high-tech firms have a higher cash ratio than old-economy manufacturing firms. This suggest that the increase in cash ratio can be due to the fact that there is an observed increase in high-tech firms, Table 6, which have a higher ratio as shown in Figure 3. However, the figure also illustrates that after the Dot-Com bubble in 2000 the cash ratio of high-tech firm dropped. This shows a negative time trend where the cash ratio of old-economy manufacturing firms show a positive time trend. In a regression on a constant and time both variables show a positive and statistically significant value for the full sample. Insert Table 8 near here However, a regression on the subsample that ranges from 2000 till 2010 shows a statistically significant cash ratio coefficient that is positive for old-economy manufacturing firms and negative for high-tech firms. This result does not provide support to the possibility that the cash ratio increased because of an higher proportion of high-tech firms in my sample. It suggest that the increased cash ratio is caused by old-economy manufacturing firms that are increasing their cash holdings and ratio in the last years. BKS also finds a positive and

15 Cash Holdings of European Firms 15 statistically significant time trend for their sample, but they find that the increase in cash ratio is higher for old-economy manufacturing firms than for high-technology firms. Their findings are in line with my results as presented above. 4. Empirical analysis As shown in Table 3 from the previous paragraph, where all firms are clustered by year, there is an statistically significant positive time trend in cash holdings over time. BKS also finds this positive trend for their U.S. sample. I start by testing if the hypothesis, of increasing cash holdings over time, also will hold for an un-clustered European sample. I therefore state the following hypothesis: Hypothesis 1: Cash holdings of European firms increase over time in the sample period ( ). To test if this hypothesis is true I use the following model: ash atio ear (1) where CashRatio is the measure of cash and marketable securities divided by total assets. The represents the constant given by the regression. The coefficient gives the interaction with the ear dummy which is 0 for the year 1989 and continuous till 21 for the year If the coefficient is significantly higher than zero, the hypothesis is accepted and there is an average increase of cash holdings over time for European firms. In addition to the finding of an positive time trend in Table 3, I also observed that after a crisis the average cash ratio dropped for the next two crisis years. This after the Dot-Com Bubble 8 ( ) and the Credit Crisis 9 ( ). There are several news articles providing the story that banks were reluctant to lend out money after the crisis. 10 This freeze up is also discussed by the FED chairman Ben S. Bernanke in his address to students in If credit markets freeze because of an crisis, firms will be more dependent on own cash reserves as obtaining a credit has become more difficult. Another reason is that firms generate less cash flow during a crisis so they have less cash to accumulate and that there are less positive NPV project available so they hold less cash as they do not need to invest. 8 Jorn Madslien, Dotcom bubble burst: 10 years ago, BBC News, March 9, Egbert Kalse, Kredietcrisis in vijf stappen, NRC, September 17, For example: Richard Tyler, Research shows banks remain reluctant to lend, The Telegraph, 27 April

16 16 Finance Master Thesis This leads me to state the following hypothesis: Hypothesis 2: Cash holdings of European firms decrease after a crisis. To test if this hypothesis is true I use model (1) and add a dummy for crisis: ash atio ear risis ummy (2) where CrisisDummy is a dummy that takes the value of 1 when a firm year is 2001, 2002, 2007 or 2008 and 0 otherwise. The hypothesis is excepted if the coefficient for the CrisisDummy is negative and statistically significant different from zero. The next part is the transaction motive for holding cash. As shown in Mulligan (1997) there is evidence for economies of scale with respect to the transaction motive. Therefore larger firms should hold less cash and gives the next hypothesis: Hypothesis 3: Larger firms hold relatively less cash than smaller firms To test this hypothesis I use the following model: ash atio i e (3) where i e is the natural logarithm of book value of assets. If the coefficient is significantly lower than zero, the hypothesis is accepted. Then there is an average decrease in cash ratio when the size of a firm increases. A firms capital expenditures also relates to the transaction motive. The expectation of the relationship of capital expenditures to cash ratio can be positive and negative. When you consider capital expenditures as a proxy for investment opportunities you expect it to be positively related to cash. The negative relationship is when you consider assets obtained trough capital expenditures to create collateral. Collateral can be used to increase a company s debt capacity and therefore reduce it cash demand. BKS finds evidence for the negative relation between capital expenditures and cash ratio for the U.S.. This leads me to state the following hypothesis: Hypothesis 4: Firms with larger capital expenditures hold relatively less cash than firm with less capital expenditures The next model is used to test this hypothesis: ash atio (4)

17 Cash Holdings of European Firms 17 where CapitalExpenditure is the ratio of a firms capital expenditure divided by book value of total assets during any given year. The hypothesis is excepted when the coefficient is negative and statistically significant. I now proceed to the precautionary motive for firms to hold cash. When firms have a negative net income it is more likely that they will be financially constrained, compared to firms with a positive net income. When a firm is financially constrained it will find it more difficult to obtain external financing and therefore hold more cash reserves as a precautionary motive. This brings the following hypothesis: Hypothesis 5: Firms with negative net income hold more cash than firms with nonnegative net income The following model I used to test this proposition: ash atio egative et ncome ummy (5) where NegativeNetIncomeDummy is a dummy that takes the value of 1 when a firm has a negative net income in a given year, and 0 otherwise. BKS finds that firms with a negative net income almost triple their cash ratio during their sample period of U.S. firms. I therefore expect this hypothesis to hold and to see therefore a positively statistically significant coefficient in the results. The market to book ratio of firms can be used as a proxy for the investment opportunities a firm has. Being short of cash would therefore have greater cost for these firms as they will have to pass on these opportunities. One would therefore expect, holding all else equal, that firms with higher market to book ratios hold more cash as a precautionary motive. I therefore state the following hypothesis: Hypothesis 6: Firms with high market-to-book ratios hold more cash than firms with a smaller market to book ratio. To test this hypothesis I use the following model: ash atio ar et-to- oo (6) where Market-to-Book is the ratio obtained by subtracting book value of equity from book value of assets and adding market capitalization and this all divided by the book value of assets.

18 18 Finance Master Thesis The next variable that this paper researches is dividends. Whether a company pays dividends or not can be seen from two perspectives. In light of the agency motive and the precautionary motive. BKS consider in their paper firms that do not pay dividends as financially constrained companies. They find therefore support for the precautionary motive to increase cash holdings for non-dividend paying firms. The free cash flow theory by Jensen (1986) suggest that firms that do not pay dividends and have poor growth opportunities accumulate more cash. From the agency motive this means that the entrenched firms accumulate cash instead of distributing it to their shareholders. These motives give the following hypothesis: Hypothesis 7: Non-dividend paying firms hold more cash than dividend paying firms. To test this hypothesis I use the following model: ash atio ividend ummy (7) where DividendDummy is a dummy that takes the value of 1 when a firm does not pay dividend in a given year and 0 otherwise. A positive and statistically significant coefficient would support the hypothesis. The agency motive is the third motive for firms to accumulate cash. The free cash flow theory of Jensen (1986) explains that firms with entrenched management would accumulate more cash than firms with less entrenched management. Corporate governance is a proxy for management entrenchment. This is because through corporate governance and shareholder rights the shareholders can take action against exploitation from entrenched management. This gives the following hypothesis: Hypothesis 8: A countries economic model, corporate governance and shareholder rights are negatively correlated with cash holdings of firms. To test this hypothesis I first look at the economic models used by countries. As shown in the previous paragraph, the countries in our sample have either an Anglo-Saxon economic model or a Continental Capitalism economic model. The first is based on free-market thinking and considered more liberal than the second. In this free-market and more liberal economy, companies will be better monitored reducing the chance on, and reducing the power of, entrenched management. I use the following model to test whether the economic model has an effect on the cash ratio: ash atio ng o- a on ummy (8)

19 Cash Holdings of European Firms 19 where Anglo-SaxonDummy is a dummy that takes the value of 1 when a firm is incorporated in the United Kingdom or Ireland (Anglo-Saxon economies). A negative coefficient would support the hypothesis stated above. Secondly, I look at shareholders rights to test the hypothesis. I use Table 2 from the paper by La Porta, Lopez-de-Silanes, Schleifer and Vishny (2002) for this. This table provides an analysis of anti-director rights for a large range of countries. The only missing country from my sample is Luxembourg, which therefore is dropped for this test. Anti-director rights is a proxy for shareholders rights. This gives me the following model for the testing the hypothesis: ash atio nti- irector ights ummy (9) where Anti-DirectorRightsDummy is a dummy that varies between 0 and 5 depending on a countries score on anti-director right. For the hypothesis to hold, the result should yield a negative coefficient. I also add some control variables to the regression. These are cash flow to assets, net working capital to assets, and leverage. The first is expected to have a positive coefficient as firms with higher cash flows can accumulate more cash. The second is a substitute for cash holdings and therefore a negative coefficient is expected. The third can be negative as well as positive. The negative coefficient argument is based on debt as a constraining factor to the firm and therefore cash will be used to reduce a firms leverage. However the hedging argument by Acharya, Almeida and Campello (2007) suggest a positive coefficient between leverage and cash holdings. Additional to the cash ratio as the dependent variable I will also conduct all the regressions on the logarithm of cash to net assets ratio. Using the logarithm as dependent variable extreme outliers can be reduced. The cash to net assets ratio subtracts cash from assets so that cash is divided by assets net of cash. 5. Results The results from the regression can be found in Table 8 and Table 9. The dependent variable of Table 8 is cash ratio and for Table 9 it is the natural logarithm of cash to net assets ratio. Columns 1 to 9 of the tables represent the results of the ordinary least squared regression on the individual hypothesis as stated in the previous paragraph. Column 10

20 20 Finance Master Thesis includes all variables used in the regression of column 3 to 9 and three control variables are added. Insert Table 8 and Table 9 near here Both tables show a positive and statistically significant correlation of cash holdings and time. These findings are consistent with the finding of BKS, that firms increased their cash holdings over time. In the third paragraph I already pointed out an increase in cash holdings over time when all observations were clustered by year. However, I also observed that during crisis years cash holdings dropped. Adding a dummy to test this yields a negative and statistically significant coefficient for a firms cash holdings during crisis years. The data therefore suggest that firms increased their cash holdings over time but during crisis years they used this cash for operations. The estimated coefficient of Size and Capex are both negative and statistically significant in the individual regressions as well as the total regression and for both dependent variables. Both variables test the transaction motive of firms to hold cash. The first shows that larger firms hold less cash than smaller firms. This is line with the finding of economies of scale in regard of the transaction motive as stated in Mulligan (1997). The second shows that the cash ratio of firms on average decreases when its capital expenditures increase. The transaction motive relates this to the fact that capital expenditures can be used as collateral and therefore the need of holding cash decreases as excess to capital markets is easier. Both coefficients are negative as expected and smaller than those in BKS. I have a negative coefficient for Size of and for Capex, where BKS finds a coefficient of for Size and for Capex, with cash ratio as the dependent variable. This suggests that firms incorporated in the U.S. experience the transaction motive more than those in the European Union. I next turn to the variables of the precautionary motive, Negative Net Income and Marketto-Book. Both variables are positive and statistically significant in the individual regressions for both dependent variables. However the dummy for negative net income is not statistically significant in the total regression with cash ratio as dependent variable and is a negative significant coefficient in the total regression with natural logarithm of cash to net assets ratio as the dependent variable. This is not what is expected as I show that when clustered by year the cash ratio of firm with a negative net income almost tripled and therefore expected for this variable to hold in the regression. The variable that gives the market to book ratio is also positive and statistically significant for the total regression with both dependent variables.

21 Cash Holdings of European Firms 21 This provides evidence that firms with a higher market to book ratio on average hold more cash and therefore have a higher cash ratio. This is in line with expectations. The estimated coefficient for the dividend dummy variable is positive and statistically significant for the individual regression as well as the total regression for both the dependent variables. It shows that firms that do not pay dividends hold more cash than those that do pay dividends. The cash ratio of a firm increases on average, holding all else equal, with 3.88% when a firm does not pay dividends. BKS show an decrease of 4% in cash ratio when a firm pays dividend and is close to our coefficient. I next look at the variables that relate to agency motive, Anglo-Saxon and Corporate Governance. I expected a negative coefficient for both variables with regard to the agency motive. However I find positive and statistically significant coefficients for both variables in the individual regression for both dependent variables. But these coefficients change to a negative value that is statistically significant in the total regressions. These coefficient provide evidence that a countries economic model and corporate governance rights are negatively correlated to cash holdings and therefore out light the agency problem in firms. Being incorporated in the United Kingdom or Ireland decreases the cash ratio of the firm with an average, ceteris paribus, of 1.61%. The corporate governance coefficient shows an decrease in cash ratio of 0.46% on average, all else equal, for every 1 change in the dummy with a range of 0 to 5 for a countries corporate governance. With respect to the control variables, I find that all three variables, Cash Flow, NWC and Leverage, are negatively correlated to cash holdings and statistically significant. The first variable suggest a positive coefficient as firm with higher cash flows could accumulate more cash. However the results show a statistically significant negative result regarding the cash flow of a firm. This negative result is also reported in the paper by BKS for the cash flow to assets variable. The second control variable is expected to be negative and therefore the result is in line with the expectation. The last control variable, Leverage, can be expected to be positive (hedging argument) as well as negative (constraining factor). I find evidence for the later as the results show a negative coefficient for both dependent variables. 6. Conclusion I documented an increase in the average cash ratio from 1989 through 2010 for European firms. The increase in the average cash ratio is smaller for the European sample than for the U.S. sample as observed in BKS. The theoretical literature provides three classes of motives

22 22 Finance Master Thesis why firms would hold cash. The first is based on the fact that firms need cash for operations and to finance its operations, the transaction motive. The second looks at the uncertainty of the future and to withstand unexpected shocks, precautionary motive. The third and last is the agency motive, this is based on agency problems within the firm where entrenched management accumulates cash for own personal benefits. Several academic papers investigated the increase in cash holdings and the determinants of this increase. Also in financial media this increase is reported as firms keep up a high cash balance for the U.S. as well as the firms in the European union. This paper contributes to the empirical literature by studying the cash holdings of European firms, where most if not all research is based on U.S. firms. By using this information I was able to research whether the European firms experience the same increase in cash holdings as the U.S. firms and whether they have the same factors driving this increase. Additional to this, there is also in sample variance between countries within the European union. Using this variance I was able to explore the role of a countries corporate governance on cash holdings of firms. In this paper I present results in support of the three motives for firms to hold cash. I started off by finding that the average cash ratio and natural logarithm of cash to net assets increased over time. Next, I found evidence that cash holdings in firms decrease during crisis years suggesting that in a crisis firms use their cash holdings to finance themselves. In support of the transaction motive I found that size and the capital expenditure ratio are negatively correlated to cash ratio. Larger firms and firms with higher capital expenditures hold less cash. My analysis finds limited support for the precautionary motive. The market to book ratio is positively correlated to cash ratio but the negative net income dummy is not statistically significant different from zero. For the agency motive I find a negative relation between a having a Anglo-Saxon based economy and the rating of its corporate governance. Firms in countries where shareholders rights are better preserved hold less cash. These results see agency problems as an indicator of accumulating cash in firms. The natural logarithm of cash to net assets as the dependent variables finds the same results as the cash ratio except for negative net income dummy. The negative net income dummy is not statistically significant different from zero for cash ratio as for the natural logarithm of cash to net assets it is a statistically significant negative value. This is not in line with expectation for the precautionary motive which would be a positive correlation. Besides this contradicting coefficient, the model has a higher R² and therefore explains more variability.