• No results found

CASH HOLDINGS OF MULTINATIONALS AND EMPLOYMENT PROTECTION

N/A
N/A
Protected

Academic year: 2021

Share "CASH HOLDINGS OF MULTINATIONALS AND EMPLOYMENT PROTECTION"

Copied!
58
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

1

CASH HOLDINGS OF MULTINATIONALS AND

EMPLOYMENT PROTECTION

MSc International Financial Management

Faculty of Economics and Business

University of Groningen

This study investigates the effect of employment protections on the relationship between internationalization and cash holdings for the period 1990 – 2013. Using differences-in-differences estimations that exploit the changes in employment protection legislation across 21 developed countries, we find that a higher level of internationalization decreases the level of cash holdings. We further find that a decrease in employment protection regulations causes multinationals to hold lower cash compared to domestic firms. On average, multinationals can better adjust themselves against such employment shocks. The findings also show that a higher level of internationalization leads to a decrease in the level of cash in countries with a stronger employment protection summary index. These results increase our understanding on differences in cash holdings between multinationals and domestic firms by employing the role of employment protection in cash management.

Classification: F23, J80, G15, G30

Keywords: Cash Holdings, Multinationals, Employment Protection, International Capital

Markets

8 JUNE 2018

BY: SIL BESSELINK

STUDENT NUMBER: S2569337

SUPERVISOR: DR. H. GONENC

CO-ASSESSOR: DR. V. ANGELINI

(2)

1

2

1. Introduction

As the world becomes more and more globalized and firms rapidly start investing overseas, the relationship between internationalization and cash holdings becomes increasingly

important. It may therefore not be surprising that a large body of literature examined the link between cash holdings and internationalization (Fernandes & Gonenc, 2016; Duchin, 2010, Pinkowitz et al., 2012). To this day, there is no clear consensus on the underlying reasons of this relationship. Firms have various reasons to hold much cash, such as financing research & development (R&D) expenditures, buffer creation or tax considerations (Foley et al., 2007; Bates et al., 2009). Very recently, Pinkowitz et al. (2016) find that R&D expenditures drive the increasing trend in cash holdings. This supports the importance of the precautionary motive explanation, which implies that firms build up their cash savings now in order to create a buffer to hedge against shocks and/or illiquidity risks in the future.

In this thesis, we aim to extend our understanding of corporate cash policy of multinationals (MNCs) by studying the implications of changes in employment protection legislation (EPL) on the relationship between cash holdings and internationalization. The idea behind using EPL is that MNCs, or firms with a higher level of internationalization, will react to the reforms differently than domestic companies (DCs) do. MNCs have sufficient capabilities to resist or adjust themselves to such reforms, while DCs do not have these capabilities.

However, the challenge in investigating the relationship between cash holdings and the level of internationalization, or between DCs and MNCs, lies in the endogenous nature of

geographic diversification of MNCs. The level of internationalization across countries is subject to constraints and depends on several firm- and country characteristics. To address the causality effect among cash holdings, geographical- and industrial diversification and

(3)

1

3 autoregressive GMM estimation. Pinkowitz et al. (2016) use matching techniques to find similarities between foreign firms and firms from the United States (US) based on major determinants of cash holdings. Thus, any differences in cash holdings between the US and foreign firms cannot be attributed to these characteristics. In this thesis, two proxies (EPL indicator and EPL summary index) are used to measure the level of employment protection legislation. For both EPL indicator and EPL summary index a sophisticated approach is used to address the endogeneity problem. To tackle the endogeneity problem for EPL indicator, a differences-in-differences (DID) estimation is exploited by comparing cash holdings of MNCs with DCs before and after the reform to identify the differences in cash holdings between MNCs and DCs. In this study, the employment reform is a truly exogenous event and is therefore issued to compare the treatment group affected by the intervention and the control group which is not affected. Further, we adjust for the panel data nature by including firm fixed effects. For EPL summary index, lagged explanatory variables are used to confront challenges and to solve endogeneity problems.

This thesis elaborates on the growing literature of finance and labor economics. A recent study of Dessaint et al. (2017) focus on the importance of employment protection reforms as a motive in corporate takeovers. Their findings show that employment protection can explain much of the differences in the activity of the takeover market around the world, and

(4)

1

4 protection as a driven force for the relation between internationalization and cash holdings. In this thesis, we investigate the cash policy of MNCs and DCs towards the employment

protection levels by using a sample covering more than 220,000 firm/year observations from 21 developed countries around the world over the period 1990-2013.

First, this thesis examines the relationship between internationalization and cash holdings. However, the direction of this relationship is still ambiguous. On one hand, both

precautionary motive and agency view can explain the positive relationship. A higher level of internationalization leads to an increase in agency conflicts, which forces managers to hold more cash. On the contrary, internationalization and diversification result in greater access to financial markets and the opportunity to transfer cash among subsidiaries. This can eventually mitigate the amount of firms’ cash holdings. Next, this study will explore if the level of employment protection has a significant impact on firms’ cash holdings. Existing theories argue both a positive and negative effect, where precautionary savings lead to a higher level of cash holdings. Stronger employment protection imposes costs on employers and restricts their flexibility, which all induces greater cash holdings. The bargaining device view explains the negative effect, which states that firms decrease the level of cash to limit the appropriated rents of employees. However, whether the increase in employment protection will have adverse effects on the cash holdings of MNCs compared to DCs is uncertain and has not been investigated before. Therefore, we investigate whether the effect of employment protection will be substantially different for cash holdings of DCs compared to that of MNCs.

(5)

1

5 influenced by the degree of employment protection. No evidence is found to support our expectations that employment protection reforms have a significantly effect on firms´ cash holdings. However, this study does provide evidence that firms in countries with stronger employment summary index, which is another indicator of employment protection, hold higher cash reserves. Moreover, the findings in this study suggest that employment protection indeed has adverse effects for firms with a different level of internationalization. We find that a decrease in employment protection leads to significantly lower cash holdings of MNCs compared to DCs. Our results also show that a higher level of internationalization leads to a decrease in the level of cash holdings in countries with stronger employment protection summary index. We argue that financial flexibility, which refers to a firms’ ability to access and restructure financing at low costs (Gamba & Triantis, 2008), plays a role in explaining our results. MNCs have characteristics that lead to higher financial flexibility and greater capital market access. This will cause MNCs to retain less on their internal funds, whereas DCs need to hold higher level of cash to hedge against future risks (Gao & Chou, 2015; Almeida et al., 2004). Since stronger employment protection is associated with greater future risks and limited access to external financing, MNCs are able to manage such risks more efficiently than DCs are.

(6)

1

6

2. Theoretical Background

2.1. Corporate cash holdings

According to Keynes (1936), there are two major benefits of cash holdings. First, firms can save transaction costs using cash to make payments, described as the transaction costs motive. Second, a firm can reserve cash to hedge against risks in the future, which is known as the incentive of precautionary savings. However, if firms have unrestricted access to external capital, there is no need to safeguard against future risks or investments.

Transaction costs, agency problems, costs of financial distress and asymmetric information are all proof that capital market imperfections exist within an economy. The presence of these financing frictions causes the costs of external finance to increase, which leads to firms investing less than the first-best optimum, eventually decreasing firm value (Fazzari et al., 1988). One way to mitigate these effects for financially constraint firms, i.e. firm with high costs of external finance, is to rely more on internal financial resources, such as cash reserves. The availability of cash enables firms to make investments without requiring access to

external capital markets, which also reduces the likelihood of financial distress costs. Also, financial flexibility might be a link in capital structure decisions as it increases access to low-cost external capital (DeAngelo & DeAngelo, 2007). It allows firms to avoid low-costs of

financial distress and alleviate the underinvestment problem.

Considering financing flexibility, financial constraints are an important issue which accounts for the variation in cash reserves. For instance, Opler et al. (1999) study the effect of

(7)

1

7 they will be able to keep investing when external funds are too expensive or when cash flows are too low. Further, Almeida et al. (2004) argue, using five alternative approaches of

financial constraints, that financial constraint firms have significantly higher cash holdings compared to unconstrained firms. Besides, they provide additional evidence that cash holding patterns vary over business cycles. The authors state that financially constrained firms raise their propensity to retain cash after a negative macroeconomic shock, whereas financially unconstrained firms do not. Related to internal financial resources, Bates et al. (2009) and Ozkan & Ozkan (2004) examine the negative effect of leverage on firms’ cash holdings. They argue that payments to debt holders reduce the opportunity to build up cash reserves.

Early studies have focused on the pros and cons of the variation in the level of cash and discuss the optimal level of cash (Opler et al., 1999; Harford, 1999; Pinkowitz et al., 2006; Denis & Sibilkov, 2009). Since cash is a primary source of financial flexibility, a large strand of literature has investigated other factors that explain differences in firms’ cash holdings. For instance, Opler et al. (1999) state that firms with strong growth opportunities have relatively greater cash holdings. In addition, smaller firms’ tend to have higher cash holdings as well. Also, a firm that pays dividends tends to hold lower amount of cash compared to a firm that does not pay dividends. Dividend behavior is related to firms’ retention policies to hold cash. Firms that pay out dividends to its shareholders are more able to raise liquidity when needed by cutting dividends (Ozkan & Ozkan, 2004; Opler et al., 2001).

(8)

1

8 illiquidity, which decreases firms’ risk and reduce retained cash policies (Titman & Wessels, 1988). The theoretical findings with regard to profitability are positive. Companies with greater profits can build up their liquidity and are more able to pay dividends (Opler et al., 1999; Dittmar et al., 2003; Ferreira & Vilela, 2004).

Not only firm, but also country level characteristics can explain differences in firms’ cash holdings. The agency costs theory is a good example to illustrate this. As stated by Opler et al. (1999), agency costs arise when the interest of shareholders differs from the interest of

managers. Managers may hold excess cash to pursue their own objectives at shareholders expense. The conflict of interest between managers and shareholders is especially likely to arise when the organization generates substantial cash. The challenge here is to motivate managers to payout shareholders rather than investing available cash below the costs of capital, retaining cash or wasting it on inefficient investments (Jensen, 1986). However, shareholders in better protected environments can force management to reduce the level of cash holdings. Therefore, firms located in countries with stronger shareholder protection tend to hold lower amount of cash (Guney et al., 2007; Dittmar et al., 2003).

2.2. Internationalization and cash holdings

In the literature, theories about the relationship between internationalization and cash holdings are seen as an ambiguous topic. One could argue that the level of internationalization will have a negative effect on firms’ cash holdings. For instance, Fernandes & Gonenc (2016) study the relationship between industrial- and geographical diversification and the level of cash. Their findings demonstrate a negative relationship, indicating that diversification leads to lower cash holdings. The major argument here is that economies of scale in cash

(9)

1

9 firms are able to cross-divisional transfer cash among subsidiaries. Duchin (2010) states that industrial diversified firms in the US have almost half of the average cash holdings compared to stand-alone firms. Furthermore, on the positive side (in terms of cash holdings), one of the benefits of corporate diversification is that expanding overseas improves access to capital markets and lowers the cost of capital (Gao & Chou, 2015). Firms can access the global capital market through a variety of different international activities, such as foreign direct investment (FDI) and mergers & acquisitions (M&A). Access to global capital markets allows firms to decrease its costs of capital due to a reduction of investors’ risk. Hence, MNCs have greater access to capital market and are therefore better positioned to secure finance for risky R&D projects and corporate headquarters can better plan firms’ cash needs. This indicates greater financial flexibility and as a consequence MNCs tend to hold lower amounts of cash (Opler et al., 1999).

On the contrary, both precautionary motive and agency view can explain the positive

relationship between internationalization and cash holdings. Many scholars argue that MNCs have greater R&D expenditures and therefore build up higher cash reserves to be able to reduce illiquidity risk in the future. For instance, Bae & Noh, (2001) show that R&D expenditures are, on average, significantly greater for MNCs than for DCs. Pinkowitz et al. (2016) show that increases in cash holdings occur all around the world, but US firms hold more cash when R&D expenditures are high. Besides, Bates et al. (2009) argue that the greater importance of R&D relative to capital expenditures has a permanent effect on cash ratios. R&D expenditures are more expensive to finance, which results in greater cash holdings.

(10)

1

10 corporate headquarters and divisional managers, which can eventually lead to inefficient resource allocation (Gao & Chou, 2015). For instance, potential agency costs may arise. (Divisional) managers and shareholders tend to view costs and benefits of cash holdings differently. Managers tend to have a greater preference for cash because it diminishes firms risk and increases their discretion. This greater preference for cash can lead managers to place too much emphasis on the precautionary motive for holding cash. Moreover, if divisional managers behave as rent-seeking agents, misallocate corporate resources or convert cash to their private benefits, potential agency conflicts arise (Rajan et al., 2000; Scharfstein & Stein, 2000). One would therefore expect that as firms become more diversified, more rent-seeking divisional managers arise, and this is associated with greater cash holdings.

In light of the relationships described above, it can be argued that internationalization can both positively or negatively affect cash holdings. Based on this, the following hypotheses will be tested:

Hypothesis 1a: Internationalization positively affects firms’ cash holdings.

Hypothesis 1b: Internationalization negatively affects firms’ cash holdings.

2.3. The effect of employment protection on the relationship between cash holdings and internationalization

Employment protection legislation is forthcoming to ensure workers against temporary productivity shocks. The academic literature has already studied different effects of employment protection on cash holdings, but the direction of this relationship seems

(11)

1

11 A large body of literature has established that changes in employment protection are

associated with many different costs, such as firing-, hiring- and agency costs. It is well-known that firing costs arise when employment protection increases. Increased firing costs may affect hiring decisions and restrict firms from freely adjusting their labor according to the demand (Lazear, 1990; Saint-Paul, 1997; Mortensen & Pissarides, 1994). However, firing costs will not occur immediately and as a consequence the precautionary demand for cash may be high for firm’s operating in high employment-protection environments (Karpuz et al., 2017). In turn, hiring new employees also becomes riskier. Millan et al. (2013) argue that it becomes more difficult for a firm to hire employees when employment protection becomes stricter, even when the quality of a new hire is exceptional. Besides, when the new hire does not perform as well as expected, the costs of dismissing will be relatively high. Other studies indicate that weaker employment protection leads to an increase in knowledge spillovers, as well as productivity growth (Bertola, 1994; Power & Lundmark, 2004). This productivity growth is based on a firms’ possibility to retain valuable workers and dismiss less valuables ones at a cheaper cost. As a consequence, firms’ can better adjust labor input, which increases flexibility and productivity growth.

Not only firing and hiring costs, but also higher agency costs will arise. John et al. (2015) argue that stronger employee protection results in greater employee-shareholder agency conflicts. Shareholders want to maximize value, but employees want higher wages and job security. This conflict would be more pronounced when employees are backed by stronger labor rights, which would result in increased agency costs. As argued by, for instance Dittmar et al. (2003) cash holdings increase when agency costs are high.

(12)

1

12 structures. Lenders are more concerned with the increase in borrower’s default risk due to restrictions on firms’ flexibility to adjust labor input. As a result, firms have limited access to external financing and cash holdings will therefore be higher among firms. To summarize, previous literature indicates that stronger employment protection imposes costs on employers and restrict firms’ flexibility. Therefore, one could expect that in order to bear the higher costs firms will maintain greater level of cash holdings.

Precautionary savings and capital market access can thus cause for a higher level of cash holdings as a result of stronger employment protection. On the other hand, one can argue that employment protection will negatively affect cash holdings. Bronars & Deere (1991) and Matsa (2010) provide evidence for this using the bargaining device view, which states that cash operates as a strategic tool in order to limit the appropriated rents of employees. Firms change the level of cash to respond to the unionization of its employees. A higher level of unionization increases bargaining power of employees in wage negotiations. Every year, unions and managers effectively bargain over the distribution of future cash flows. Before these meetings, managers can decrease the amount of available cash by, for example, dividend payouts. The decrease in available cash provides firms with an incentive to counteract the bargain and avoid increases in wages. Simplified, firm with greater cash holdings would find it more difficult to refuse labor union’s demand for a pay rise. Therefore, the bargaining device view indicates that an increase in labor protection may lead to a

decrease in the level of cash holdings, in order to prevent a possible pay rise from happening.

Based on the theories discussed above, the following hypothesis will be tested:

Hypothesis 2a: Employment protection is positively related with the level of cash holdings.

(13)

1

13 According to the Organization for Economic Co-operation and Development (OECD) annual Going for Growth report (2017), the governments in most countries need employment reforms to tackle the low-growth trap and prepare for the coming technological changes. However, employment reforms are challenging and to ensure that all people can benefit, additional measures are needed. Many countries follow the OECD recommendations to increase job creation by, for instance, lowering labor tax wedges on low-wage workers. The employment reform varies per country, but each government must take advantage of the synergies between labor, product and financial market reforms. However, to do take advantage of the synergies, each government must be able to identify the pros and cons for all different kind of firms. So far, there has been no systematic evidence if the effect of employment protection on cash holdings differs for MNCs compared to DCs.

Building on theories described above, it could be argued that MNCs have different

characteristics compared to DCs to adjust themselves against macroeconomic shocks. We argue that financial flexibility plays an important role on the effect of employment protection as a driven force for the relationship between internationalization and cash holdings. The reasoning behind this thought is that MNCs have greater financial flexibility and capital market access, which cause MNCs to retain less on their internal funds, whereas DCs need to hold higher level of cash to hedge against future risks. As an increase in employment

protection is associated with greater future risks and limited access to external financing, it can therefore be argued that employment protection reforms have a substantially lower impact on MNCs’ cash holdings compared to the cash holdings of DCs. Further, Almeida et al (2004) argue that financially constrained firms increase cash holdings after a negative

(14)

1

14 strengthens our theory is based on the study of Duchin (2010), who states that MNCs are able to transfer cash among subsidiaries. This increases their financial flexibility and could

improve MNCs ability to cope with employment reforms and manage risks more efficiently than DCs can. Therefore, it can be argued that employment protection will weaken

(strengthen) the positive (negative) relationship between internationalization and cash

holdings. In light of the relationships described above, the following hypothesis will be tested.

Hypothesis 3: the effect of employment protection will weaken (strengthen) the positive (negative) relationship between internationalization and cash holdings.

3. Data and variables

3.1 Sample and data sources

The analysis covers 21 developed countries for which data on major employment protection reforms are available. EPL data is available until 2013; therefore, the sample covers the period from 1990 to 2013. The analyses of previous studies cover the period 1985-2007 (Dessaint et al.,2017; Karpuz et al., 2017). They stop in 2007 and argue that the global

financial crisis had a substantial impact on both firms’ labor law, cash holdings and it affected economies differently. By extending the sample period to 2013, this analysis will verify if results stay consistent by including the crisis period. Our main data source is Compustat North America and Global. Financial firms (SIC 6000-6999), utilities (4900-4999) and government-related sectors (9000-9999) are excluded from the analysis. These requirements result in a final sample consisting of 19,580 firms with in total 220,077 observations.

(15)

1

15 level variables are collected from various sources. GDP per capita is drawn from World Bank. Firm-level fundamental characteristics are collected from Compustat North America and Global. Foreign sales and foreign assets are drawn from Datastream. Sources for each variable can be found in Appendix A, Table 1.

3.2. Research design and variables

One focus of this thesis is to examine the relationship between the level of internationalization and cash holdings. The dependent variable cash holdings is measured by the ratio of cash and short-term investments to total assets [cash ratio]. Additionally, cash ratio will be replaced by net cash ratio (the ratio of cash divided by net assets) as a robustness check in this study.

There is no standard approach for measuring the degree of internationalization. However, the most common approach to measure the degree of internationalization is based on foreign sales ratio; foreign sales as a percent of total sales (Pinkowitz et al, 2016; Fernandes & Gonenc, 2016). In this study the ratio of foreign sales to total sales [multinationality] is used. To test the results for robustness, this study includes foreign assets ratio (foreign assets to total assets) as a proxy for multinationality.

(16)

1

16 requirements on firing, conditions of the use of temporary contracts etcetera. For a detailed description of the reforms, We refer to Appendix B of Simintzi et al. (2015).

In total, 17 major labor reforms are identified in the sample period 1990-2013. Of the 17 reforms, five lead to an increase in employment protection (Belgium, 1998; France, 1990, 1993; Italy, 1991; Switzerland, 1994), whereas eleven lead to a decrease in employment protection (Australia, 2005; Austria, 2003; Denmark, 1999; Germany, 1997; Italy, 1998; the Netherlands, 1999; Norway, 1994; Portugal, 1991; Spain, 1994 and Sweden, 1993, 1997). Some countries have two major reforms (France, Italy and Sweden). We also include countries in our sample that did not have any such employment protection reforms (Canada, Finland, Greece, Ireland, Japan, New Zealand, United Kingdom (UK) and the US). These countries are always in the control group.

In order to create the EPL indicator and to measure employment protection, we include a variable EPLk,t that is set to zero for all countries (k) in the beginning year (t), which is 1990

(EPLk,t = 0). In each of the subsequent years the value of EPLk,t remains constant if there is no

major employment protection reforms in county (k) in year (t). EPLk,t will increase by one if

there is a major reform increasing employment protection in country (k) in year (t) (Rk,t = +1),

where Rk,t refers to the employment protection reform. EPLk,t will take the value of minus one

if employment protection went down in country (k) in year (t) (Rk,t = -1). This EPLk,t variable

captures large, long-run changes in employment regulation over time and treats all

employment protection reforms equally. For any given country (k) in year (t), the following equation is applied:

(17)

1

17 The variable EPLk,t is not comparable across countries, but it is a good indicator over time

within a country.

In addition, both EPLpositiveindicator and EPLnegative indicator variables are created. These variables examine separately positive and negative changes to employment protection.

EPLpositive k,t increases by one whether a major reform increases employment protection during

the year (t) in the target firm country (k), whereas EPLnegative k,t decreases by one if a major

reform decreases employment protection during the year (t) in the target firm country (k). For instance, Sweden has major reforms in 1993 and 1997 that decrease employment protection, which implies that EPLnegativeindicatorwill take the value of minus two as off 1997. In each of the subsequent years the value of EPLnegativeindicator remains constant as there were no other major employment protection reforms in Sweden (country) as of 1997 (t). The variable EPLnegative k,t is set to zero for all countries in which no major reform that decreases

employment protection take place.

In this study we use EPL summary index, identified in the study of Karpuz et al. (2017), as another proxy to measure employment protection. EPL summary index is calculated by unweighted average of version-1-sub-indicator for temporary- and regular contracts reported by the OECD. Each year, OECD publishes EPL indices for each member country by

(18)

1

18

3.3. Empirical model

To test the hypotheses, we first use EPL indicator, EPLpositive indicator and EPLnegative

indicator as a proxy for employment protection legislation. We exploit intertemporal variations in labor regulation across 21 developed countries using a DID research design to examine the impact of employment protection legislation on the relationship between cash holdings and internationalization. The main assumption in a DID design is that of parallel trends, which implies that the outcome variable would have behaved in a similar way across treated groups and control groups (absent treatment). For this analysis, it implies that cash holdings would have developed in a similar way across treated groups including labor reforms and control groups in the absence of labor reforms. To tackle the endogeneity problem, we add the exogenous variable EPL indicator to the regression analysis to compare cash holdings of MNCs with DCs before and after the reform to identify the differences between MNCs and DCs. Second, we will test our hypotheses using EPL summary index. In order to control for the endogeneity problem here, lagged variables are included in the regression analysis. To exploit within group variation over time, firm fixed effects are also included.

In order to test the first hypotheses, whether the level of internationalization affects firms’ cash holdings, the following regression specification by ordinary least squares (OLS) is used:

Where i denotes firm, k denotes country and t is year. CASHi,t is the cash ratio of firm i in

time period t, MUL is the level of internationalization in firm i in time period t.

(19)

1

19 variables are divided into two groups, respectively firm- and country characteristics. The variable Xi,t contains firm-level characteristics based on extant studies (Karpuz et al., 2017,

Fernandes & Gonenc, 2016). We include size, Tobin’s Q, leverage ratio, tangibility, net working capital, cash flow, capital expenditures, R&D expenditures, net equity issues and net debt issues. In the analyses, Tobin’s Q is calculated as the sum of book value of total assets and market value of equity minus book value of equity divided by book value of total assets. Furthermore, size is measured by taking the natural logarithm of book value of assets in USD. Cash flow, measured by net income plus depreciation divided by book value of total assets, is included as a control variable as previous literature find that firms with higher cash flows hold higher cash reserves (Bates et al., 2009; Fernandes & Gonenc, 2016). Further, dividend is included as it is found that firms who pay dividends to its shareholders are more able to raise funds when needed by cutting dividend and hold therefore lower cash reserves (Opler et al., 1999). To measure dividend, a dummy variable is included that takes the value of one if a firm pays dividend and zero otherwise.

The variable Zk,t is included to control for different country level characteristics. To ensure

that the regression outcome is not driven by trade differences and to control for different economic conditions the log of GDP per capita is included. Creditor rights (Djankov et al., 2007) and disclosure requirements index (La Porta et al., 2006) are also included to control for differences in investor protection. Firm and year fixed effects (fi) are included in all

analyses to capture other omitted variables correlated with cash holdings. All variables are defined in Appendix A, Table 1.

To test for the second hypotheses, whether cash holdings are significantly positively or

(20)

1

20 including EPLk,t, which refers to employment protection reforms. This leads to the following

regression model:

Different regressions will be run including EPL indicator, EPLpositive indicator, EPLnegative indicator and EPL summary index as proxies for EPKk,t. The variable Xi,t and Zk,t contain

identical firm and country level characteristics as in regression equation (2).

The following regression model is utilized to test the third hypothesis, which state that employment protection has a significantly different effect for cash holdings of MNCs compared to DCs:

In equation (4) an interaction variable between the level of internationalization and

employment protection is included (MULi,t * EPLk,t). Equation (4) comprises similar firm and

country level control variables as in both equation (2) and equation (3).

(21)

1 21 4. Empirical results 4.1. Summary statistics

Table 1 reports the summary statistics for the variables used in this study, which comprises 220,077 observations in 21 developed countries including 19,580 firms. We note that summary statistics of the variables are consistent with other studies. For instance, the mean value of variable multinationality in our sample is 17.77%, which is in line with 18.63% found in the study of Fernandes & Gonenc (2016). Further, the cash ratio mean of 18.41% in our sample is in line with the cash ratio found by Fernandes & Gonenc (2016). The standard deviation of cash ratio (0.2011) is quite large, meaning a large variation in the variable. A similar conclusion can be drawn when looking at the standard deviation of net cash ratio (0.7787). This is also visible in the minimum and maximum values of both cash ratio and net cash ratio. Based on the minimum (-2.00) and maximum value (2.00) of EPL indicator, we note that there are countries with both two major reforms that increases and decreases employment protection. This is consistent with the reforms identified in our sample. Firms included in the sample obtained on average a Tobin’s Q of 1.86 and a leverage ratio of 0.44. In the sample, there are some firms that are completely leveraged. Further, we see from the summary statistics that the median firm did not change its R&D expenditures and its payout of dividends. This could indicate that the median firm does not have any dividend payouts or R&D expenditures. Average R&D expenditures in the sample are approximately 3.2%. Further, we note that on average firms have 27.87% tangible assets of total assets. Some firms do not acquire any tangible assets (0.00%), whereas others have 91.96% tangible assets of total assets. This shows a large diversity.

(22)

1

22 dependent variables EPLpositiveindicator and EPL indicator and EPLnegative indicator and EPL indicator are in line with the expectations. It will not form an issue since both variables are used in separate regressions.

Table 3 shows summary statistics of the main variables EPL indicator, EPL summary index, cash ratio, multinationality (foreign sales ratio) and foreign assets ratio per country. Table 3 also reports the country split in the total number of observations and the amount of EPL major reforms per country which is used to create EPL indicator. In the sample, US have most firm/year observations followed by Japan, UK, Australia and Germany. On average, total sales from firms in Austria, Finland, Ireland, the Netherlands and Switzerland comprise for more than 40% of foreign sales.

(23)

1

23

Table 1, Summary statistics

This table reports summary statistics for all variables in our analysis from the period 1990 till 2013. The data are taken from 21 developed countries. All variables, except cash flow, EPL summary index, EPL indicator, EPLpositive indicator, EPLnegative indicator, dividend, creditor rights, disclosure, GDP per capita are winsorized at 1% level in both tails. The variable cash flow is winsorized at a 5% level in both tails. Table 1 presents number of observations (N), the mean, minimum (Min), maximum (Max) and standard deviation (Std. Dev.) of variables for the full sample. Capital expenditures, cash flow, dividend, leverage, net debt issues, net equity issues, net working capital, tangibility, R&D expenditures, size and Tobin’s Q represent the firm characteristics control variables. Creditor rights, disclosure, log (GDP per capita) represent the country level characteristics. Definitions and sources for each variable can be found in Appendix A, Table 1.

Sample descriptive statistics

Variable N Mean Min Max Median Std. Dev.

Cash ratio 220,077 0.1841 0.0000 0.9258 0.1140 0.2011

EPL indicator 220,077 0.2340 -2.0000 2.0000 0.0000 0.8465 EPLpositive indicator 220,077 0.1128 0.0000 2.0000 0.0000 0.4355 EPLnegative indicator 220,077 -0.1844 -2.0000 0.0000 0.0000 0.4415 EPL summary index 220,077 1.1330 0.2533 4.1042 1.1220 0.8378 Multinationality 220,077 0.1777 0.0000 1.0000 0.0000 0.2801 Foreign assets ratio 118,922 0.1386 0.0000 0.9918 0.0002 0.2280 Net cash ratio 220,077 0.2943 0.0000 6.4270 0.0940 0.7787

Firm level control variables

Leverage 220,077 0.4420 0.0138 1.0000 0.4380 0.2378

Net working capital 220,077 0.0536 -1.4870 0.6546 0.0532 0.2617 R&D expenditures 220,077 0.0322 0.0000 0.6222 0.0000 0.0878 Size 220,077 5.1662 -0.6812 10.4380 5.1922 2.1656 Tobin’s Q 220,077 1.8630 0.4368 15.0933 1.2393 2.0403 Capital expenditures 220,077 0.0499 0.0000 0.4071 0.0281 0.0683 Cash flow 220,077 0.0039 -0.5911 0.1997 0.0506 0.1795 Dividend 220,077 0.4912 0.0000 1.0000 0.0000 0.4999

Net debt issues 220,077 0.1052 -0.3861 0.7871 0.0555 0.1605 Net equity issues 220,077 0.0664 -0.1024 1.2614 0.0000 0.2057 Tangibility 220,077 0.2787 0.0000 0.9196 0.2293 0.2287

Country level control variables

Creditor rights 220,077 1.8762 0.0000 4.0000 2.0000 1.0882

Disclosure 220,077 0.7948 0.2500 1.0000 0.7500 0.1790

(24)

24

Table 2, Panel B Correlation matrix

This table presents the correlation matrixes of the variables employment in the regression. Panel A presents the correlation coefficients between all firm level employed variables. Panel B presents the correlation between all country level control variables.

Panel A

Cash ratio

EPL indicator EPLpositive indicator EPLnegative indicator EPL summary index Multi nationality Leverage Net working capital R&D expenditures Size Tobin’s Q Cash flow Capital expen-ditures Dividend Net debt issues Net equity issues Tangi- bility Cash ratio 1.000 EPL indicator -0.043 1.000 EPLpositive indicator -0.040 0.740 1.000 EPLnegative indicator -0.024 0.749 0.110 1.000

EPL summary index -0.116 0.099 0.501 -0.347 1.000

Multinationality -0.056 0.026 0.120 -0.080 0.120 1.000

Leverage -0.372 0.105 0.090 0.066 0.137 0.010 1.000

Net working capital 0.092 0.039 0.033 0.025 -0.007 0.079 -0.479 1.000

R&D expenditures 0.375 0.000 -0.042 0.041 -0.188 0.039 -0.041 -0.060 1.000 Size -0.301 0.124 0.035 0.149 0.092 0.278 0.139 0.128 -0.225 1.000 Tobin’s Q 0.323 -0.056 -0.053 -0.030 -0.174 -0.029 0.008 -0.249 0.360 -0.297 1.000 Cash flow -0.303 0.087 0.061 0.068 0.098 0.110 -0.049 0.289 -0.414 0.447 -0.317 1.000 Capital expenditures -0.084 -0.066 -0.027 -0.071 -0.108 -0.015 -0.092 -0.095 -0.056 -0.044 0.074 -0.005 1.000 Dividend -0.183 -0.005 -0.126 0.117 0.032 0.008 0.044 0.066 -0.175 0.393 -0.156 0.310 -0.085 1.000

Net debt issues -0.224 0.018 0.043 -0.016 0.100 0.057 0.497 -0.121 -0.078 0.226 -0.041 0.032 0.038 0.058 1.000

Net equity issues 0.381 -0.102 -0.051 -0.101 -0.100 -0.087 -0.145 -0.133 0.279 -0.359 0.418 -0.472 0.133 -0.214 -0.116 1.000

Tangibility -0.365 -0.040 -0.087 0.027 -0.020 -0.067 0.006 -0.188 -0.217 0.160 -0.152 0.118 0.461 0.101 0.226 -0.095 1.000

Panel B

Creditorrights

Log(GDP per

capita) Disclosure Creditorrights 1.000

Log(GDP per capita) -0.034 1.000

(25)

1

25

Table 3 Summary statistics for the EPL indicator, EPL summary index, cash ratio and the level of internationalization per country

This table shows country summary statistics of the main variables used in the regression. Column 2 presents the total number observations per country (N). Column 3 and Column 4 report the mean and standard deviation (Std. Dev.) of EPL indicator variable. Column 4 and column 5 present respectively the amount of reforms increasing employment protection (EPLpositive) and the amount of reforms decreasing employment protection (EPLnegative), based on Siminitzi et al. (2015). Columns 6-15 report the mean, median and standard deviation (Std. Dev.) of EPL summary indicator, cash ratio and multinationality per country. Column 16-18 reports number of observations (N), mean and standard deviation (Std. Dev) of foreign assets ratio per country. Definitions of the variables can be found in Appendix A, Table 1.

EPL indicator EPL summary index Cash ratio Multinationality Foreign assets ratio

Country N Mean Std. Dev EPLpositive EPLnegative Mean Median Std. Dev Mean Median Std. Dev Mean Median Std. Dev N Mean Std. Dev.

(26)

1

26

Table 4 Mean and count per year of main variables

This table presents number of observations (N) and mean value of the variables EPL summary index, cash ratio, multinationality used in the regressions per year. Column (6) and (7) represent number of observations (N) and the mean value of foreign assets ratio per year. Cash ratio is the dependent variable. Sources and calculations of

each variable can be found in Appendix A, Table 1. EPL sum mary index Cash ratio Multin

ationality Foreign assets ratio

year N Mean Mean Mean N Mean

(27)

1

27 In this study, we split EPL indicator variable in both EPLpositive indicator and EPLnegative

indicator to examine separately positive and negative changes to employment protection. To do so, we have to verify whether both EPLpositiveindicator and EPLnegativeindicator are significant different from zero. With a DID research design for both EPLpositiveindicator and EPLnegativeindicator, we examine whether employment protection reforms have a significantly different effect between cash holdings of MNCs compared to DCs. In our DID research design, employment protection is our exogenous treatment variable. We create a dummy variable (Neg. EPL dummy) which takes the value of one if a reform decreases employment protection (treatment = 1), and zero when there is no employment protection reform

(treatment = 0). Next, we create a dummy variable to identify the group exposed to the treatment. In our study, we split the sample group in MNCs and DCs using MNC dummy, which takes the value of one if foreign sales are greater than 25% of total sales in a given year, and zero otherwise. This is in line with the dummy used in the study of Fernandes & Gonenc (2016). In our sample, 27.22% of the firms are identified as MNCs (companies with foreign sales greater than 25% of total sales).

Table 5, Panel A shows the DID estimation outcome with cash holdings as the dependent variable. We take the differences in the mean of cash ratio between treated firms (MNCs) and control firms (DCs) before and after the employment reform to compare cash holdings of MNCs with DCs. We note that a negative employment reform increases the mean of cash holdings of domestic firms from 0.1878 to 0.2403 on average, whereas a negative

employment reform decreases the mean of cash holdings from 0.1708 to 0.1343. Further, Panel A of Table 5 reports that the differences in mean of cash ratios between MNCs and DCs after a negative employment reform are statistically significant different from zero at -8.90%, which is particularly large. These descriptive statistics confirm that a statistically significant difference between cash holdings of MNCs compared to DCs after a decrease in employment protection exists, which confirms our third hypothesis.

(28)

1

28

Table 5 Differences-in-differences estimation of Neg. EPL dummy and Pos. EPL dummy variable

This table shows in Panel A the differences-in-differences (DID) estimation of negative EPL dummy (Neg. EPL Dummy). The sample period is 1990-2013. The dependent variable is cash ratio, measured by cash and short term investment divided by book value of total assets. Neg. EPL dummy takes the value of one when reform decreases employment protection, and zero otherwise. MNC dummy takes the value of one if foreign sales to total sales are greater than 25%, and zero otherwise. Panel B shows DID estimation of positive EPL dummy (Pos. EPL Dummy), which takes the value of one when reform increases employment protection, and zero otherwise. The symbols ***, **, * denote statistical significance at the 1%, 5%, and 10% levels, respectively. Δ refers to the delta between treated and control groups

4.2. The relationship between the level of

internationalization and cash holdings

We begin by investigating the relationship between cash holdings and internationalization. Table 6 reports the results of our baseline regression obtained by pooled time-series cross-sectional regressions of the cash ratio as presented in equation 2. In this model, the degree of internationalization is calculated as the ratio of foreign sales to total sales [multinationality]. Table 6 presents the results and consists of seven models.

Panel A MNC Dummy

Neg EPL Dummy 0 1 Δ

0 0.1878 0.1708 -0.0170***

1 0.2403 0.1343 -0.1060***

Δ 0.0525*** -0.0365***

DID estimation -0.0890***

Panel B MNC Dummy

Pos EPL Dummy 0 1 Δ

0 0.1961 0.1665 -0.0296

1 0.1685 0.1425 -0.0260

Δ -0.0276 -0.0240

(29)

1

29 We follow previous research in defining firm characteristics as control variables (Fernandes & Gonenc, 2016; Pinkowitz et al., 2016; Duchin, 2010). We also incorporate firm fixed effects and year fixed effects in the regressions. Column (2) of Table 6 presents the results of multinationality regressions, and this model is used to test the first hypothesis. Interestingly, the coefficient of multinationality is -0.017 and statistically significant at a 1% level. This result generally confirms that a higher level of internationalization decreases cash holdings, which is in line with the findings of Fernandes & Gonenc (2016). Based on the argument mentioned in their study, we argue that economics of scale in cash management exist.

Column (1) of Table 6 presents the results including five firm control variables – leverage, net working capital, R&D expenditures, size and Tobin’s Q – that explain firms’ cash holdings. The control variables tend to have the predicted effects on cash holdings in our estimation results. For instance, both size and Tobin’s Q appear to be significant predictors of changes in cash holdings, where size has a negative effect and Tobin’s Q a positive effect. As previous research recognized, smaller firms tend to have higher cash holdings compared to larger firms and firms with higher profits can build up their liquidity which implies greater cash holdings (Opler et al., 1999). Moreover, the coefficient of leverage is statistically significant at 1% level and negatively related to cash holdings, which is consistent with theoretical predictions.

(30)

1

30 coefficient of tangibility is significantly negative (-0.415) in Column (5), suggesting that more tangible assets reduces firms’ cash holdings. While all other firm level control variables stay consistent after including tangibility, the coefficient of net working capital turns negative. This is interesting, but, it is however possible that more fixed assets lead to a lower need of cash reserves. Tangible assets can be sold if the need for cash arises, leading to management utilizing its assets in a more efficient, which can result in a greater amount of net working capital (Fernandes, 2011). This is also in line with the trade off theory, which state that an inverse relationship between cash holdings and net working capital exists. For instance, Bates et al. (2009) and Opler et al. (1999) found that firms can only maintain high level of net working capital or cash, indicating the negative relationship.

Finally, to examine the effects of country characteristics, Column (7) of Table 6 documents the regression analysis including country level and firm level variables together. Results reveal that country level characteristics also explain differences in firms’ cash holdings. Creditor rights and disclosure have both significantly negative coefficients, suggesting that companies in better protected countries with more disclosure decline the amount of cash. This is consistent with theoretical predictions of Ferreir & Vilela (2004). Similarly, the coefficient of GDP per capita has the predicted effect on cash holdings in our results. The coefficient is 0.039, implying that countries with higher economic performance increases cash holdings.

(31)

1

(32)

1

32

Table 6: Multinationality and cash holdings

This table shows the full-sample results of the OLS regression analyses for the period 1990-2013. The dependent variable is cash ratio representing cash plus short-term investment dividend by book value of total assets. Multinationality refers to the degree of internationalization, when value is missing, it is set to zero. R&D expenditures refers to research & development expenditures, which is set to zero when missing. Tangibility refers to firms’ asset tangibility. Definitions and sources of all variables are given in Appendix A, Table 1. Year and firm fixed effects are included in all regressions. A dataset of 19,580 firms is used including 220,077 observations. The symbols ***, **, * denote statistical significance at the 1%, 5% and 10% levels, respectively. Robust, clustered standard errors are provided in the brackets.

Dependent variable Cash ratio

(1) (2) (3) (4) (5) (6) (7)

Multinationality -0.017*** -0.018*** -0.014*** -0.017*** -0.015*** -0.015*** [0.003] [0.003] [0.003] [0.003] [0.003] [0.003] Leverage -0.209*** -0.209*** -0.217*** -0.212*** -0.206*** -0.233*** -0.235***

[0.005] [0.005] [0.005] [0.006] [0.005] [0.006] [0.006] Net working capital 0.020*** 0.020*** 0.021*** 0.008 -0.013*** -0.030*** -0.029***

[0.005] [0.005] [0.005] [0.005] [0.005] [0.005] [0.005] R&D expenditures -0.107*** -0.105*** -0.114*** -0.096*** -0.060*** -0.055*** -0.057*** [0.016] [0.016] [0.016] [0.015] [0.015] [0.014] [0.014] Size -0.023*** -0.023*** -0.022*** -0.021*** -0.018*** -0.017*** -0.019*** [0.001] [0.001] [0.001] [0.001] [0.001] [0.001] [0.001] Tobin’s Q 0.011*** 0.011*** 0.012*** 0.007*** 0.009*** 0.006*** 0.005*** [0.000] [0.000] [0.000] [0.000] [0.000] [0.000] [0.000] Capital expenditures -0.281*** -0.086*** -0.080*** [0.009] [0.008] [0.008] Cash flow -0.041*** -0.010** -0.008* [0.005] [0.004] [0.004] Dividend -0.001 -0.001 -0.001 [0.001] [0.001] [0.001]

Net debt Issues 0.059*** 0.092*** 0.092***

[0.005] [0.005] [0.005]

Net equity Issues 0.168*** 0.144*** 0.145***

[0.004] [0.003] [0.003] Tangibility -0.415*** -0.389*** -0.387*** [0.007] [0.007] [0.007] Creditor rights -0.062*** [0.006] Disclosure -1.774** [0.077]

Log(GDP per capita) 0.039***

(33)

1

33 4.3. The effect of employment protection on het relationship between cash holdings and internationalization

In the previous section, we showed, based on firm fixed effects, that if firms’ degree of internationalization increases, they need to hold less cash than before. In this section, we analyze if cash holdings are significantly influenced by the degree of employment protection, and whether this effect is influenced by the level of internationalization.

First, we account for employment protection by including EPL indicator variable in our analysis. Table 7, Column (1) reports the regression results of the baseline effect of employment protection reforms on cash holdings. We estimate that the average effect of employment reforms using EPL indicator is -0.001, but insignificant at the 10% level, suggesting that cash holdings are not influenced by the degree of employment protection. Second, we examine the same relationship including separate positive and negative changes to employment protection, respectively EPLpositive indicator and EPLnegative indicator. Column (3) reports the results. The estimated coefficients are, again, not statistically significant. In summary, our OLS results show that the degree of employment protection does not have a significant effect on firms’ cash holdings.

Our next analysis examines the effect of employment protection as a driven force for the relation between internationalization and cash holdings. We control for the same set of firm and country characteristics as in the regressions reported earlier. We note that in all models the coefficient of multinationality is negative and significant, and therefore identical as reported in the regressions before. To determine whether MNCs and DCs differ significantly in terms of the extent to which employment protection affects their cash holdings, an

(34)

1

34 adverse effects for firms with a different level of internationalization. Surprisingly, the

coefficient of EPL indicator interaction is positive and significant at a 1 % level. However, the results are based on somewhat limited positive employment protection reforms. Therefore, our focus in the regression is placed on the interaction between both positive employment protection reforms and multinationality, and negative employment protection reforms and multinationality; EPLpositive interaction and EPLnegative interaction respectively. While the coefficient on the interaction between EPLpositive and multinationality is slightly positive (at 0.001) but insignificant in Column (4), the coefficient on the interaction between EPLnegative and multinationality is 0.018 and significant at a 1% level. This indicates that negative employment protections reforms do indeed influence cash holdings of MNCs differently compared to DCs. The values of EPLnegative interaction coefficient vary between minus two and zero. Therefore, a positive sign indicates that a decrease in employment protection lead to significantly lower cash holdings of MNCs compared to DCs. These findings support

hypothesis 3, which states that the effect of employment protection will strengthen the negative relationship between internationalization and cash holdings. The explanation of our main result may be that MNCs are less financial constraint than DCs. MNCs have greater access to capital markets and are more financial flexible. Our results therefore support the findings of Almeida et al. (2004), who argue that financially constraint firms raise their propensity to retain cash after a macroeconomic shock, whereas financially unconstrained firms do not.

(35)

1

35 on average, firms’ cash holdings increase after an employment reform that decreases

employment protection. Hence, there is still little evidence that employment protection

(36)

1

36

Table 7: Employment protection, cash holdings and multinationality

Dependent variable Cash ratio

(1) (2) (3) (4)

EPL indicator -0.001 -0.004

[0.003] [0.003]

EPLpositive indicator 0.005 0.004

[0.005] [0.005]

EPLnegative indicator -0.002 -0.007**

[0.003] [0.004] EPL indicator interaction 0.011***

[0.003]

EPLpositive interaction 0.001

[0.005]

EPLnegative interaction 0.018***

[0.005] Multinationality -0.015*** -0.014*** -0.015*** -0.010*** [0.003] [0.003] [0.003] [0.004] Leverage -0.235*** -0.235*** -0.235*** -0.235***

[0.006] [0.006] [0.006] [0.006] Net working capital -0.029*** -0.029*** -0.029*** -0.029***

[0.005] [0.005] [0.005] [0.005] R&D expenditures -0.057*** -0.057*** -0.057*** -0.057*** [0.014] [0.014] [0.014] [0.014] Size -0.019*** -0.019*** -0.019*** -0.019*** [0.001] [0.001] [0.001] [0.001] Tobin’s Q 0.005*** 0.005*** 0.005*** 0.005*** [0.000] [0.000] [0.000] [0.000] Capital expenditures -0.080*** -0.080*** -0.080*** -0.081*** [0.008] [0.008] [0.008] [0.008] Cash flow -0.008* -0.008* -0.008* -0.008* [0.004] [0.004] [0.004] [0.004] Dividend -0.001 -0.001 -0.001 -0.001 [0.001] [0.001] [0.001] [0.001] Net debt issues 0.093*** 0.093*** 0.093*** 0.093*** [0.005] [0.005] [0.005] [0.005] Net equity issues 0.145*** 0.145*** 0.145*** 0.145*** [0.003] [0.003] [0.003] [0.003] Tangibility -0.387*** -0.387*** -0.387*** -0.387*** [0.007] [0.007] [0.007] [0.007] Creditor rights -0.063*** -0.060*** -0.057*** -0.056*** [0.007] [0.007] [0.008] [0.008] Disclosure -1.777** -1.746** -1.722** -1.710** [0.766] [0.749] [0.752] [0.727] Log(GDP per capita) 0.039*** 0.038*** 0.038*** 0.038*** [0.004] [0.004] [0.004] [0.004] Constant 1.630*** 1.603*** 1.570*** 1.570*** [0.611] [0.598] [0.601] [0.581] R-squared 0.772 0.772 0.772 0.772 Nr. of observations 220,077 220,077 220,077 220,077

(37)

1

37 Next, using the alternative approach as described by Karpuz et al. (2017), a different measure of employment protection can be obtained by using EPL summary index. As in previous research (e.g. Karpuz. et al., 2017; Fernandes & Gonenc, 2016), we include lagged values to estimate the results. Table 8 provides the results, in which identical firm and country level characteristics are included as in previous regressions. All coefficients of the control variables show similar signs as reported earlier. Therefore we do not discuss on these variables.

Column (1) of Table 8 reveals the baseline effect of employment protection reforms on cash holdings. The coefficient of EPL summary index shows a positive, significant sign at a 1% level. This indicates that stronger employment protection leads to greater cash holdings, which supports the precautionary motive. This also provides supportive evidence for

hypothesis 2a. Next, we add the interaction between EPL summary index and multinationality in Column (2) to compare if MNCs and DCs adjust their cash holdings differently as a result of employment protection legislation [EPL summary interaction]. The coefficient of the interaction variable is -0.008 and significant at a 1% level, which suggests that on average an increase of employment protection will have a lower effect on cash holdings of MNCs compared to DCs. In other words, a higher level of internationalization decreases the level of cash in countries with stronger employment protection summary index. This again supports hypothesis 3 and shows that MNCs are better able to resist themselves against employment protection reforms than DCs. Therefore, they hold lower cash reserves.

Fixed effects

Year X X X X

(38)

1

38

Table 8: Employment protection, cash holdings and multinationality

This table shows OLS regression for determining the effect of employment protection on the relation between cash holdings and multinationality. The dependent variable is cash holdings, measured by cash and short-term investments divided by book value of total assets. Definitions and sources of all variables are given in Appendix A, Table 1. Year and firm fixed effects are included in all regressions. A dataset of 19,580 firms including 200,890 observations is used. The symbols ***, **, * denote statistical significance at the 1%, 5% and 10% levels, respectively. Robust and clustered standard errors are provided in the brackets.

Dependent variable Cash ratio

(1) (2)

EPL summary index 0.006* 0.006*

[0.003] [0.003]

EPL summary index interaction -0.008*** [0.002] [0.002] Multinationality -0.001*** -0.001*** [0.000] [0.000] Leverage -0.232*** -0.232*** [0.007] [0.007] Net working capital -0.026*** -0.026***

[0.005] [0.005] R&D expenditures -0.064*** -0.063*** [0.015] [0.015] Size -0.020*** -0.020*** [0.001] [0.001] Tobin’s Q 0.005*** 0.005*** [0.000] [0.000] Capital expenditures -0.084*** -0.084*** [0.008] [0.008] Cash flow -0.007 -0.007 [0.005] [0.005] Dividend -0.001 -0.001 [0.001] [0.001] Net debt issues 0.092*** 0.092***

[0.005] [0.005]

Net equity issues 0.148*** 0.148*** [0.003] [0.003] Tangibility -0.383*** -0.382*** [0.008] [0.008] Creditor rights 0.022*** 0.018*** [0.003] [0.003] Disclosure 0.394 0.395 [0.283] [0.253]

Log(GDP per capita) 0.045*** 0.045***

(39)

1

39 4.4. The effect of the financial crisis

Table 9 provides results, i.e. whether the global financial crisis had a substantial impact on firms’ labor law, cash holdings and economies. To investigate this, the sample is split and separate regressions are run based on sample period 1990 – 2007. If hypothesis 1b still holds, the coefficient of multinationality variable should be significantly negative. The results shown in Column (1) do provide supportive evidence for hypothesis 1b, with the coefficient of multinationality being -0.014 and significant at a 1% level.

In Column (2) of Table 9, we investigate the impact of employment protection on firms’ cash holdings by including EPL indicator. We note that our results are still insignificant, and therefore do not support hypothesis 2a or 2b. However, when including EPLpositive indicator and EPLnegative indicator, the estimated coefficient of EPLpositive in Column (4) is positive and significant at a 1% level. This indicates that tighter employment protection increases firms’ cash holdings, which supports hypothesis 2a that cash holdings are positively influenced by the degree of employment protection.

The interaction term between EPL indicator variable and multinationality in Column (3), which we have used to investigate hypothesis 3, remains to be positive and significant at a 10% level. This contradicts hypothesis 3. In Column (4) EPLpositive interaction and EPLnegative interaction are included. We note that both coefficients are insignificant, i.e. we do not find support for hypothesis 3. This could be explained by the limited number of observations, which is 151,899.

(40)

1

40 index in Column (1) is still positive and significant. This provides evidence that stricter

employment protection increase firms’ cash holdings, which is in line with hypothesis 2a. To support hypothesis 3, the coefficient of the interaction term should be negative and

(41)

1

41

Table 9: Employment protection, cash holdings and multinationality

This table shows OLS regression for determining the effect of employment protection on the relation between cash holdings and multinationality. The dependent variable is cash holdings, measured by cash and short-term investments divided by book value of total assets. Definitions and sources of all variables are given in Appendix A, Table 1. Year and firm fixed effects are used in all regressions. A dataset of 17,541 firms including 151,899 observations is used. The symbols ***, **, * denote statistical significance at the 1%, 5% and 10% levels, respectively. The sample period is from 1990 – 2007. Robust and clustered standard errors are provided in the brackets.

Dependent variable Cash ratio

(1) (2) (3) (4) (5)

EPL indicator -0.001 -0.001

[0.003] [0.003]

EPLpositive indicator 0.008* 0.003

[0.005] [0.006]

EPLnegative indicator -0.001 0.009

[0.003] [0.006] EPL indicator interaction 0.006*

[0.004]

EPLpositive interaction 0.003

[0.006]

EPLnegative interaction 0.009

[0.006] Multinationality -0.014*** -0.014*** -0.014*** -0.014*** -0.013*** [0.004] [0.004] [0.004] [0.004] [0.005] Leverage -0.217*** -0.217*** -0.217*** -0.217*** -0.217***

[0.008] [0.008] [0.008] [0.008] [0.008] Net working capital -0.025*** -0.025*** -0.025*** -0.025*** -0.025***

[0.006] [0.006] [0.006] [0.006] [0.006] R&D expenditures -0.068*** -0.068*** -0.068*** -0.068*** -0.068*** [0.017] [0.017] [0.017] [0.017] [0.017] Size -0.018*** -0.018*** -0.018*** -0.018*** -0.018*** [0.001] [0.001] [0.001] [0.001] [0.001] Tobin’s Q 0.005*** 0.005*** 0.005*** 0.005*** 0.005*** [0.000] [0.000] [0.000] [0.000] [0.000] Capital expenditures -0.088*** -0.088*** -0.088*** -0.088*** -0.088*** [0.009] [0.009] [0.009] [0.009] [0.009] Cash flow -0.005 -0.005 -0.005 -0.005 -0.005 [0.005] [0.005] [0.005] [0.005] [0.005] Dividend -0.002 -0.002 -0.002 -0.002 -0.002 [0.001] [0.001] [0.001] [0.001] [0.001] Net debt issues 0.084*** 0.084*** 0.084*** 0.084*** 0.084*** [0.005] [0.005] [0.005] [0.005] [0.005] Net equity issues 0.143*** 0.143*** 0.143*** 0.143*** 0.143*** [0.004] [0.004] [0.004] [0.004] [0.004] Tangibility -0.367*** -0.367*** -0.367*** -0.367*** -0.367*** [0.009] [0.009] [0.009] [0.009] [0.009] Creditor rights -0.061*** -0.060*** -0.059*** -0.053*** -0.053*** [0.002] [0.003] [0.003] [0.005] [0.005] Disclosure 0.000 0.000 0.000 0.000 0.000 [0.000] [0.000] [0.000] [0.000] [0.000]

Log(GDP per capita) 0.057*** 0.057*** 0.057*** 0.057*** 0.057*** [0.004] [0.004] [0.004] [0.004] [0.004]

Constant 0.015 0.010 0.010 -0.003 -0.001

[0.042] [0.042] [0.042] [0.043] [0.043]

(42)
(43)

1

43

Table 10: Employment protection, cash holdings and multinationality

This table shows OLS regression for determining the effect of employment protection on the relation between cash holdings and multinationality. The dependent variable is cash holdings, measured by cash and short-term investments divided by book value of total assets. Definitions and sources of all variables are given in Appendix A, Table 1. Year and firm fixed effects are used in all regressions. A dataset of 17,541 firms including 134,630 observations is used. The symbols ***, **, * denote statistical significance at the 1%, 5% and 10% levels, respectively. The sample period is from 1990 – 2007. Robust and clustered standard errors are provided in the brackets.

Dependent variable Cash ratio

(1) (2)

EPL summary index 0.019*** 0.020***

[0.003] [0.003]

EPL summary index interaction -0.008***

[0.002]

Multinationality -0.001*** -0.001***

[0.000] [0.000]

Leverage -0.212*** -0.211*** [0.008] [0.008] Net working capital -0.024*** -0.024***

[0.007] [0.007] R&D expenditures -0.071*** -0.070*** [0.018] [0.018] Size -0.019*** -0.018*** [0.002] [0.002] Tobin’s Q 0.005*** 0.005*** [0.001] [0.001] Capital expenditures -0.087*** -0.086*** [0.010] [0.010] Cash flow -0.006 -0.006 [0.006] [0.006] Dividend -0.001 -0.001 [0.001] [0.001] Net debt issues 0.084*** 0.084***

[0.006] [0.006]

Net equity issues 0.145*** 0.145*** [0.004] [0.004] Tangibility -0.366*** -0.366*** [0.009] [0.009] Creditor rights 0.023*** 0.019*** [0.002] [0.002] Disclosure 0.000 0.000 [0.000] [0.000]

Log(GDP per capita) 0.050*** 0.050***

(44)

1

44 4.5. Robustness checks

The outcomes reported in Table 6, Table 7 and Table 8 may also be the result of alternative explanations. As a robustness check, this study investigates two of these possible alternatives. First, we include a different approach to measure the level of cash holdings. The same control variables are issued to check whether the results stay consistent. Table 11 presents the results using net cash ratio as the dependent variable. Net cash ratio is calculated as cash plus short-term investments divided by total assets minus cash and short-short-term investments.

Table 11 shows that a higher level of internationalization reduces cash holdings, supporting hypothesis 1b. Overall, the use of net cash ratio does not differ our regression outcomes much. However, Column (2) shows that the coefficient of EPL indicator is -0.067 and significant, suggesting that stronger employment protection reforms decreases firms’ cash holdings. The same conclusion can be drawn when looking at EPLpositive interaction and EPLnegative interaction in Column (3) and EPL summary index in Column (4). This confirms hypothesis 2b that employment protection negatively affects firms’ cash holdings. No differences can be found when looking at EPLnegative interaction and EPL summary index interaction variable in Column (3) and (4). This indicates that MNCs can better adjust themselves against employment shocks, which supports hypothesis 3.

(45)

1

45 Table 12 presents the results using EPL summary index interaction. This coefficient is still negative and significant, which is consistent with our previous findings.

Referenties

GERELATEERDE DOCUMENTEN

Therefore in situations of high uncertainty where information asymmetries are increased, as measured by higher cash flow volatility or higher R&D expenses, Continental

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

Hypothesis 2: The corporate cash holdings of relatively small listed firms in countries in countries with good shareholder rights are only lower than poor shareholder

effects on the cash holdings of MNCs is uncertain and is, to our knowledge, not investigated before. For that purpose, we a particular interested in examining

Despite the previous notions, a research from Pinkowitz and Williamson (2001) concludes that US and German companies hold lower amounts of cash compared to Japanese ones. This is

26 Interpreting the corruption coefficient leads to the following impact of corruption on cash holdings: If the control of corruption estimate of a host-country

The underlying assumption of this hypothesis is based on the existence of foreign operations that enable the tax planning strategies (Foley.. 17 et al., 2007) Hence, there

Investment size, is the log of R&D expenditures, i.e., log(rd) Strong FTR, is based on the nationality of CFO and CEO and a binary variable that indicates whether their