Control and Cash Holdings within Corporate Groups –
Empirical Evidence from UK parent firms and their Subsidiaries
By Matthias Jelden
Abstract
This thesis investigates monitoring and control problems and their effects on cash holdings within corporate groups using a sample of 896 listed UK parent firms and their 2.934 domestic and 1.013 EU-located subsidiaries. Control within the corporate group is measured by the degree of multinational activity, degree of equity ownership in subsidiaries and the difficulty by which the corporate group can control related or unrelated and focused or diversified subsidiaries. The empirical findings of this thesis provide support for the agency theory applied to cash holdings and for a tax cost explanation of higher consolidated cash holdings of multinational corporations. Finally, the findings show that higher controlling stakes of a parent firm in a subsidiary prevent subsidiary managers from building up cash.
Key words: cash holdings, control, agency costs, multinational corporations, MNCs, subsidiaries
University of Groningen, Faculty of Economics and Business University of Uppsala, Faculty of Economics & Business
Master Thesis
Table of Contents
1. Introduction 3
2. Literature review 5
2.1. Optimal levels of cash 5
2.2. Cash holdings, monitoring and agency costs 6
2.3. Cash holdings and multinational activity 7
2.4. Control within MNCs 8
2.5. Control and ownership 11
2.6. Control and relatedness 11
2.7. Control and diversification 13
3. Hypotheses development 14
3.1. Degree of multinationality and cash holdings 14
3.2. Degree of equity ownership in subsidiaries and cash holdings 15
3.3. Relatedness and cash holdings 16
3.4. Diversification and cash holdings 17
4. Methodology 19
4.1 Operationalization of variables 19
4.1.1. Operationalization of variables for the corporate group regressions 19 4.1.2. Operationalization of variables for the subsidiary regressions 21
4.2. Firm-specific determinants of cash holdings 22
4.3. Model description and justification 24
5. Data description 26
5.1. Sample selection 26
6. Empirical results 28
6.1. Results of the corporate group regressions 28
6.2. Results of the subsidiary regressions 32
6.3 Robustness tests 34
7. Conclusion 35
7.1. Suggestions for future research 37
8. References 38
1. Introduction
A large body of corporate finance literature finds that agency theory can explain cash ratios of firms. The basic idea is that a positive relationship between corporate cash holdings and agency problems exists (Jensen and Meckling, 1976; Jensen, 1986). Agency problems occur when imperfect monitoring by shareholders enables managers to build up cash holdings. With large cash holdings, they are not subject to external capital market discipline and enjoy more flexibility and discretion in their
investment decisions. However, empirical results on the relationship between agency costs and level of cash holdings are still inconclusive. Some studies find that entrenched managers build up cash (Dittmar et al. 2003; Kalcheva and Lins, 2007) while other scholars state that entrenched managers spend excess cash quickly e.g. on acquisitions (Harford et. al., 2007). Other scholars find no evidence (Opler et al., 1999; Mikkelson and Partch, 2003) that supports this relationship.
Later studies analyse the agency theory applied to cash holdings in focused and diversified firms (Tong, 2008) or private and public firms (Von Eije, 2012; Brav, 2009). However, little empirical evidence regarding agency costs of cash in domestic and multinational corporations (hereafter called MNCs) has been provided so far. This is surprising, since both firm types face corporate governance differences which can affect cash holdings. In this thesis, I focus on control and agency problems occurring within corporate groups and examine how they affect the corporation’s consolidated cash holdings and the subsidiary cash holdings.
I study control within corporate groups on three different dimensions. First, I examine monitoring and control problems associated with multinational activity. Second, I analyse control problems associated with equity ownership in subsidiaries. Finally, I test whether specific subsidiary characteristics such as being unrelated (instead of related) and diversified (instead of focused) complicate monitoring by the parent firm. For these four measures I analyse their impact on cash holdings of corporate groups and subsidiaries.
This thesis contributes to existing literature in several ways. First, it sheds light on governance problems within corporate groups. Second, it provides evidence that the agency theory applied to cash holdings (Jensen and Meckling, 1976; Jensen, 1986) affects parent firms differently from subsidiaries. In particular, the location of the subsidiary proves to be highly relevant for consolidated cash holdings in parent firms while a higher equity ownership in the subsidiary reduces cash holdings of the subsidiary.
2. Literature review 2.1. Optimal levels of cash
If capital markets are perfect, assuming absence of taxes, agency costs, asymmetric information and bankruptcy costs, a firm will be indifferent to holding either cash reserves or finance operations with external funds (Modigliani and Miller, 1958). Since the costs of raising new cash and the opportunity costs of holding cash are irrelevant here, the cash policy of a firm neither increases nor destroys shareholder value. However, the prediction of this model when compared to economic reality with imperfect capital markets seems to lack explanatory power. In this setting, a firm’s cash policy can affect shareholder value since marginal benefits and marginal costs of holding cash exist. According to Opler et al. (1999), marginal costs associated with holding cash comprise tax disadvantages and a liquidity premium due to a lower rate of return compared to other illiquid assets. Marginal benefits of holding cash comprise a lower probability of financial distress, a reduction in costs for raising external capital and minimization of the underinvestment problem (Ferreira and Vilela, 2004). Opler et al. (1999) propose a trade-off model by equaling marginal benefits and marginal costs of cash to derive an optimal level of cash holdings which increases shareholder value. This model is depicted in Figure 1.
The optimal level of cash holdings (A*) is determined by the intersection of the marginal costs and marginal benefits curve. At this level of cash holdings, firm managers maximize the value of the firm. Opler et al. (1999) assume that the marginal costs of holding cash remains constant and does not vary with the level of cash holdings. In contrast, the marginal benefits curve is negatively related to the level of cash. The marginal benefits of holding cash can be regarded as the marginal costs of a cash shortage, which decrease as the level of cash holdings increases by one additional unit. This indicates that an additional unit in cash holdings reduces the probability of a cash shortage and thereby a reduction in marginal costs of a cash shortage.
2.2. Cash holdings, monitoring and agency costs
Applying these empirical findings to the trade-off model in Figure 1, imperfect monitoring by shareholders would induce entrenched managers to build up cash, leading to a shift from an optimal level of cash holdings A* along the marginal benefits curve to a sub-optimal level of cash holdings A-. In this setting, the marginal benefits curve and marginal costs curve do not intersect, indicating a disequilibrium or sub-optimal level of cash. Here, the marginal costs of holding cash exceed the marginal benefits leading to a reduction in shareholder wealth.
Contrary to the empirical findings presented above, Harford et al. (2007) study US firms and find that firms under weak corporate governance structures have lower cash ratios. Their analysis supports the view that in poorly governed US firms, cash is not hoarded but spent quickly by managers on acquisitions and capital expenditures. Opler et al. (1999) and Mikkelson and Partch (2003) find no supporting evidence that high cash firms are not effective in monitoring managers or incentivizing managers to align their interests with shareholders. Instead of an agency cost explanation for higher cash holdings, Mikkelson and Partch (2003) find that high cash reserves support growth and minimize costly external financing.
2.3. Cash holdings and multinational activity
In comparison to purely domestic companies, firms that operate internationally may have additional motives to hold cash. Existing literature points out two main motives that apply here – a tax motive and a precautionary motive due to greater and more volatile investment opportunities abroad. Foley et al. (2007) analyses cash holdings of US multinationals under tax considerations. They show that firms facing high tax costs on repatriating foreign income hold more cash in their foreign subsidiaries. However, these higher cash holdings in the subsidiaries are not offset by lower cash holdings of the parent firm leading to higher consolidated cash holdings. Their estimates show an increase in consolidated cash holdings (defined as cash/net assets) by 7.9% as tax costs on
Ramirez and Tadesse (2009) examine the effect of culture (uncertainty avoidance) and
multinationality on consolidated cash holdings of corporations. According to their empirical findings, firms located in countries with high uncertainty avoidance hold more cash. A high score indicates that a society tends to rely on established norms, rules and routines to avoid uncertainty associated with an unknown future. This finding supports the precautionary motive to hold more cash as a natural hedge. Further evidence indicates a positive relationship between the degree of
internationalization (measured by foreign sales ratio) and the level of cash holdings. Ramirez and Tadesse (2009) argue that multinational firms tend to hold more cash than domestic firms due to tax costs associated with repatriating foreign income and a longer business cycle. Besides they state that the business cycles of MNCs are more unpredictable which requires larger cash holdings.
2.4. Control within MNCs
A large body of literature addresses specific monitoring problems (and agency costs) that can be associated with multinational activity and therefore present challenges for MNCs rather than purely domestic corporations. According to Ghoshal & Bartlett (1990) multinational corporations can be characterized as an enterprise of geographically dispersed and goal-disparate organizations. MNCs face pressures for global integration and pressures for local responsiveness (Prahalad & Doz, 1987). Global integration refers to global deployment of resources and pursuing strategic objectives at the corporate group level by integrating activities across countries. In contrast, local responsiveness requires strategic decision making within a local context (Roth et al., 1990).
Frequently, many host governments require that MNCs employ local managers when entering the host country (Wright et al., 2002), therefore MNCs send expatriates from headquarters to the foreign subsidiary to increase local control. Consequently, managers with different nationalities and cultural backgrounds work together. Since expatriate managers and local managers hold different cultural values and beliefs, and different risk preferences, a larger cultural distance increases the probability of interpersonal conflicts of interest (Hofstede, 1980). These conflicts can occur on two dimensions, first at the subsidiary level between local and expatriate managers and second in the relationship between headquarters and subsidiary (Chang and Taylor, 1999).
According to Filatotchev and Wright (2011), both internationalization theory as well as agency theory can explain corporate governance problems in MNCs. International business literature identifies three major dimensions on which MNCs face governance problems - at the firm-level, at the national institutional level and in the internationalization process. Whereby governance issues in the internationalization process can stem from foreign direct investment (FDI), control aspects in partially-owned subsidiaries and the headquarters-subsidiary agency relationship. Besides the classical principal-agent relationship between external shareholders and managers, within MNCs there exists an additional (or extended) agency relationship between headquarters and subsidiaries (Chang and Taylor, 1999). The corporate headquarters is seen as principal and the subsidiary as agent. The MNC invests in the subsidiary and expects that it acts in the interest of the MNC. Thus, headquarters managers try to minimize agency costs by effectively controlling subsidiary managers to maximize shareholder value. Wright et al. (2002) state that conflicts between headquarters and foreign subsidiaries generate agency problems when the MNC cannot effectively monitor or control foreign subsidiaries. As a result, foreign subsidiary managers can conduct activities that enhance the managers’ private benefits but reduce shareholder wealth.
Buckley and Strange (2010) state that internal transaction costs of MNCs are higher than of domestic corporations due to the internal division of labor across borders. According to Buckley and Strange (2010), internal transaction costs comprise costs of information acquisition and transmission,
coordination costs, and costs of aligning interests of various stakeholders within the MNC. Individuals in the multinational organization have limited information processing capability and are not likely to possess all relevant information, therefore the cost of information acquisition increases. Aligning interests of stakeholders within the firm, e.g. by providing incentives, increases motivation costs or bonding costs.
Empirical studies investigate effects of internationalization decisions and generally propose a
relationship between the degree of internationalization and agency costs. Analyzing US firms, Doukas (1995) finds that higher foreign involvement is associated with higher expected agency costs. On average US public firms announcing expansion of foreign operations do not receive a positive market reaction because investors assume that in geographically diversified firms, agency problems to be more pronounced.Wright et al. (2002) find that firms with higher exposure to foreign markets face more negative valuation effects than firms with lower degrees of international business. Besides, agency costs increase with firm size indicating that large multinational firms face the highest agency costs. Also, Jiraporn et al. (2006) find that internationally-diversified firms are valued at a discount compared to domestic firms and Lee and Wyatt (1990) report a negative relationship between announcements of international joint ventures and abnormal returns of firms. Their findings support the view that self-interested managers obtain private benefits from international diversification and thereby reduce shareholder value.
Additional support for higher agency and monitoring costs of MNCs is provided by Doukas and Pantzalis (2003) by showing that in comparison to domestic firms, monitoring of MNCs with high foreign involvement is more difficult and costly for creditors. Higher foreign involvement is
2.5. Control and ownership
Filatotchev and Wright (2011) argue that control problems within MNCs can be reduced by increasing the equity ownership degree in foreign subsidiaries. Chang and Taylor (1999) add to this by showing that the power of MNCs to exert control over subsidiaries varies with the degree of equity ownership in the unit. Their findings indicate that on average a higher equity ownership in a subsidiary is related to higher output and staffing control measures. O'Donnell (2000) finds that headquarters use less monitoring, such as supervision and bureaucratic monitoring, when subsidiary autonomy is high. It is argued that autonomous subsidiaries are more difficult to monitor and therefore monitoring costs are higher compared to subsidiaries in which the parent firm has relatively higher equity stakes.
Gaur and Lu (2007) report that MNCs prefer full ownership modes over partial ownership modes to gain high controlling stakes in foreign subsidiaries, however this may not always be feasible. First, higher investment is associated with higher risk and therefore MNCs may prefer shared ownership with another local or foreign party (Chang and Taylor, 1999). Second, local governments try to avoid that foreign firms acquire controlling shares in local firms and may impose restrictions that limit the control of foreign investors in local companies (Stulz and Wasserfallen, 1995). Thus, the ability to fully control subsidiaries abroad depends on risk preferences of the MNC and governmental requirements.
2.6. Control and relatedness
market opportunities. Thus, the MNC has to manage a trade-off by ensuring subsidiary autonomy to absorb and develop knowledge to exploit local investment opportunities on the one hand, while on the other hand exerting an adequate level of control. Moreover, headquarters has to provide the subsidiary with enough discretion to determine an adequate level of assets (including cash) for utilizing local market inefficiencies (O'Donnell, 2000). In addition, the difficulty of monitoring increases when the subsidiary operates in another industry segment than the parent firm. In this case, the subsidiary’s business activities can be regarded as being unrelated (to a certain extent) from the parent firm’s business activities. A low level of relatedness between subsidiary and parent firm requires delegating decision making authority to the subsidiary managers and thereby leads to a reduction of control by headquarters managers (O'Donnell, 2000). Nohria and Ghoshal (1994) add that MNCs cannot fully centralize coordination, decision making authority and control, as
headquarters is not fully aware of the subsidiary’s local market knowledge, industry knowledge, or its specialized information. Therefore, the parent firm provides subsidiaries with sufficient discretion, decision-making autonomy and may decentralize control so that the subsidiary has more flexibility in investment decisions when new investment opportunities arise.
2.7. Control and diversification
Chang and Taylor (1999) argue that the problem of foreign subsidiary control increases when the foreign subsidiary grows in size and becomes more diversified (measured by the number of different business segments). When subsidiary size and diversification increase, information ambiguity may intensify and it becomes more difficult for headquarters to fully comprehend and verify the legitimacy of decisions made by the foreign subsidiary management.
Another stream of literature that compares agency conflicts in stand-alone firms and diversified firms generally suggests that agency problems are more likely to occur in diversified firms rather than focused firms due to a higher difficulty of control by external investors (Tong, 2008; Stulz 1990; Billett and Mauer, 2003). Jensen (1986) argues that managers use diversification to entrench themselves or increase compensation, prestige and power. Therefore, excess funds are used for diversification into segments that best serve the interest of the manager and increase his or her personal wealth.
In diversified firms, cross-subsidization can be employed by managers to shift resources from well-performing divisions with good growth prospects to low-well-performing divisions (Tong, 2008). As a result, cross-subsidization can be misused by managers when funds are shifted to divisions that serve primarily the manager’s interest and do not enhance shareholder value. This problem is assumed to be less severe in focused firms due to fewer diversification opportunities and relatively higher transparency.
3. Hypotheses development
Foley et al. (2007) show that firms that face high tax costs on repatriating foreign income hold more cash in their foreign subsidiaries which in turn leads to higher consolidated cash. Hence, repatriation taxes can be regarded as a major determinant of MNC cash holdings. To take this tax effect into account, I analyse multinational activity within the EU (27)1and outside the EU (27) separately. Tax
costs on repatriating foreign earnings are assumed to be irrelevant within the EU (27), since parent firms with subsidiaries located in other EU member states are exempted from withholding taxes on dividend flows within the corporation2.
In the following sections, the effects of the explanatory variables on cash holdings are analysed both at the corporate group level (consolidated cash) and subsidiary level (unconsolidated cash).
Correspondingly, I formulate hypotheses that apply to both levels.
3.1. Degree of multinationality and cash holdings
Chang and Taylor (1999) point out that headquarters' uncertainty about foreign subsidiary decision-making increases with geographical and cultural distance. Imperfect monitoring of foreign
subsidiaries is expected to lead to agency problems when entrenched subsidiary managers build up cash levels. These higher subsidiary cash levels are expected to be reflected also in higher
consolidated cash holdings of the corporate group. Therefore, I expect that higher multinational activity would result in higher consolidated cash holdings of the corporate group. On the contrary, I would assume that domestic subsidiaries are easier to monitor due to lower geographical distance, lower language barriers, lower communication costs and lower cultural distance. Hence, higher domestic activity is expected to lead to more effective monitoring by headquarters and less opportunities for entrenched subsidiary managers to build up cash. This leads to the following hypotheses:
1Composed of Austria, Belgium, Bulgaria, Cyprus, the Czech Republic, Denmark, Estonia, Finland, France,
Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, the Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, and the United Kingdom
2Council Directive 2003/123/EC of 22 December 2003 amending Directive 90/435/EEC on the common system
H1a: The higher the relative number of EU subsidiaries belonging to a corporate group, the higher the consolidated cash of the corporate group.
H1b: The higher the relative number of Non-EU subsidiaries belonging to a corporate group, the higher the consolidated cash of the corporate group.
H1c: Cash holdings of EU subsidiaries are higher compared to cash holdings of domestic subsidiaries. 3.2. Degree of equity ownership in subsidiaries and cash holdings
For the sake of simplicity, the following hypotheses relate to subsidiaries in general. In the
subsequent analysis, I will compare if the anticipated effects influence EU and domestic subsidiaries differently.
Since higher equity ownership in a subsidiary is expected to improve monitoring and control by a parent firm, I expect that higher equity ownership enables parent firms in limiting managerial discretion of subsidiary managers. Due to reduced managerial discretion, subsidiary managers have fewer opportunities to accumulate cash, resulting in lower cash holdings of the subsidiary and the corporate group (after consolidation of parent and subsidiary cash holdings). In partially-owned subsidiaries I would expect more agency problems since the parent firm’s controlling stakes are insufficient to effectively govern the subsidiary leading to higher cash holdings. The following hypotheses are formulated:
H2a: The higher the equity ownership in subsidiaries belonging to a corporate group, the lower the consolidated cash of the corporate group.
3.3. Relatedness and cash holdings
Wade and Gravill (2003) define parent firms and subsidiaries operating in congruent areas e.g. a similar industrial group or industry, as being related. Accordingly, subsidiaries operating in a different industrial group or industry than the parent firm can be defined as being unrelated. If the subsidiary operates in a different industry i.e. it is unrelated from the parent, monitoring by the parent firm is expected to be more complicated since it has not so much specialized knowledge about the industry in which the subsidiary operates. In contrast, monitoring of a related subsidiary is expected to be easier since both parent firm and subsidiary operate in the same industry. Since the parent firm is expected to hold specialized industry knowledge, it may be better able to estimate the cash requirements of a related subsidiary compared to an unrelated subsidiary.
I hypothesize that when the parent firm and subsidiary are active in different industries, the
subsidiary enjoys greater independence and is less subject to parent supervision. For the parent firm it may be more complicated to effectively control unrelated subsidiaries compared to related subsidiaries. Headquarters may not have the same relevant industry information and expertise which the unrelated subsidiary has. Besides, headquarters managers may have only limited access to specific industry information and may have difficulties to interpret the information correctly. Therefore, it may be difficult for headquarters managers to determine optimal cash levels of unrelated subsidiaries leaving entrenched subsidiary managers with more opportunities to accumulate cash.
I expect that information asymmetries between the parent firm and subsidiary are higher when both firms operate in different industries. Higher information asymmetries may allow entrenched
subsidiary managers to build up cash levels in unrelated subsidiaries. In contrast, information asymmetries between headquarters and related subsidiaries are expected to be relatively lower, leaving subsidiary managers with fewer opportunities to hoard cash. The following hypotheses are formulated:
H3a: The higher the relative number of related subsidiaries belonging to a corporate group, the lower the consolidated cash of the corporate group.
3.4. Diversification and cash holdings
When considering subsidiaries only, one can distinguish between subsidiaries that are active in only one industry segment or more. Where the first are frequently labeled focused (or stand-alone) firms and the latter are diversified. Tong (2008) defines firms that operate in at least two different business segments as being diversified and firms active in only one business segment as being focused.
According to his findings, higher agency conflicts occur in diversified firms compared to focused firms.
Information ambiguity between headquarters and subsidiary may intensify when a subsidiary is operating in several different industry segments. In this scenario, it becomes more difficult for headquarter managers to fully comprehend and verify the legitimacy of decisions made by subsidiary management. In comparison to focused subsidiaries, managers of diversified subsidiaries may have more opportunities to misrepresent or withhold relevant information. In this way, a sub-optimal cash level could be justified by providing headquarters management with incorrect information.
Moreover, headquarters may allocate too much cash to the subsidiary on the basis of misstated information provided by subsidiary managers or due to a lack of expertise of the diverse activities of the subsidiary. This would lead to a sub-optimal cash level (or excess cash level) too.
On the other hand, since diversified firms are active in several business segments, they generate cash flows that are imperfectly correlated. This enables diversified firms to hold relatively less cash compared to focused firms since the risk of a cash shortage is lower (Duchin, 2010). In this scenario, diversified subsidiaries would have relatively lower cash holdings compared to focused subsidiaries.
Agency problems are assumed to be less severe in focused subsidiaries since monitoring by headquarters is expected to be more effective due to lower costs of information acquisition and easier interpretation of that information. This leads to the following hypotheses:
H4a: The higher the relative number of diversified subsidiaries belonging to a corporate group, the higher the consolidated cash of the corporate group.
4. Methodology
4.1 Operationalization of variables
4.1.1. Operationalization of variables for the corporate group regressions
Table 1 outlines the variables being used in the first regression model. In this section I explain in detail, how the variables are computed and what they measure.
For the first regression model, which estimates the determinants of consolidated cash holdings of corporate groups, I compute the dependent variable LOG_CASH. This variable is the natural logarithm of cash and cash equivalents scaled by net assets, which is defined as total assets minus cash and cash equivalents. The data for LOG_CASH is retrieved for 2010 (year-end).
The independent variable EU_SUBS is the number of all recorded subsidiaries located in the EU (27) that belong to a corporate group scaled by the total number of subsidiaries that belong to this corporate group i.e. domestic subsidiaries, EU subsidiaries and Non-EU subsidiaries. Subsidiaries that are located in the UK are classified as domestic subsidiaries and therefore are not included in the number of recorded EU subsidiaries. The resulting ratio measures the extent of multinational activity of a corporate group within the EU.
These ownership variables measure the extent of control a parent firm exercises over their subsidiaries. For example, a higher ratio of wholly-owned subsidiaries compared to all subsidiaries belonging to a corporate group would indicate a higher level of control and vice versa.
The sample also contains subsidiaries (or more precisely associates) that are partially-owned i.e. equity ownership <50%. I do not incorporate these associates in the subsequent analysis since for these partially-owned firms full consolidation of accounts (in accordance with IAS 27) does not apply. IAS 273prescribes that parent firms with substantial equity ownership in subsidiaries i.e. >=50% have
to prepare consolidated financial statements in which the financial accounts of parent and subsidiary are added line by line. This applies to assets, liabilities and equity as well as to income and expenses. However for associate firms, accounting standard IAS 284applies which prescribes cost and equity
methods for consolidation.
Moreover, I include the independent variable REL_SUBS in the first regression model. This variable represents the ratio of related subsidiaries to all subsidiaries belonging to a corporate group. To compute this measure, I count the number of subsidiaries of a corporate group that operate in the same industry as the parent firm and divide it by the total number of related and unrelated subsidiaries. To identify whether the subsidiary operates in the same industry group or not, I use two-digit and three-digit US SIC primary codes. Two-digit US SIC primary codes represent the major industry group and three digit codes define the industry group more specifically. I measure
relatedness at the two-digit and three-digit level to increase the robustness of the model.
To differentiate between focused and diversified subsidiaries, I incorporate the variable DIV_SUBS. Following Tong (2008), diversified subsidiaries are defined as firms that have at least two business segments with different US SIC primary codes. Following this definition, I classify subsidiaries as either diversified or focused. Then, I count the number of diversified subsidiaries that belong to a certain corporate group and scale it by the total number of focused and diversified subsidiaries of the same corporate group. A higher ratio indicates a higher number of diversified subsidiaries that need to be monitored by the parent firm.
3IAS 27 - Consolidated and Separate Financial Statements. Retrieved January 4, 2012 from:
http://ec.europa.eu/internal_market/accounting/docs/consolidated/ias27_en.pdf
4IAS 28 - Investments in Associates. Retrieved January 4, 2012 from:
Thus, this variable relates to the level of complexity that has to be managed by the parent firm and therefore the difficulty with which it can effectively control subsidiaries’ cash holdings policies. To enhance robustness, I compute this measure based on three-digit and four-digit US SIC primary codes (Tong, 2008).
4.1.2. Operationalization of variables for the subsidiary regressions
Table 2 outlines the variables being used in the subsidiary regression model. In this section I explain how I operationalize and compute these variables.
For the second regression model, which estimates the determinants of unconsolidated cash holdings of a subsidiary, I compute the dependent variable LOG_CASH. This variable is the natural logarithm of cash and cash equivalents scaled by net assets, which is defined as total assets minus cash and cash equivalents. The data for LOG_CASH is retrieved for 2010 (year-end). To obtain a
normally-distributed and well-behaved variable, I take the natural logarithm.
The dummy variable EU_DUM takes on the value 1 if the subsidiary is located in one of the EU member countries except the UK. If it is located in the UK it takes on the value 0 and is defined as a domestic subsidiary. This measure simply indicates whether the subsidiary is foreign or domestic and is included in the model to analyse monitoring and control problems associated with
multinationality.
To operationalize the difficulty by which the parent firm can monitor and control the cash holdings of subsidiaries, I incorporate REL_DUM and DIV_DUM. The dummy variable REL_DUM indicates
whether a subsidiary is related to its parent firm by operating in the same industry or not. It takes on the value 1 if it is related and 0 if otherwise. As for the first regression model, I also measure
relatedness with US SIC primary codes at two-digit and three-digit levels.
The dummy variable DIV_DUM denotes whether a subsidiary is diversified or focused. It takes on the value 1 if it is diversified and 0 if otherwise. Following Tong (2008), I calculate this measure based on three-digit and four-digit US SIC primary codes. Additionally, I include the variable #DIFF_BUS which represents a subsidiaries’ total number of different business segments.
4.2. Firm-specific determinants of cash holdings
Existing literature points out important firm-specific determinants that explain the variance in corporate cash holdings. These determinants are used as control variables in the regression models. Ferreira and Vilela (2004) find that investment opportunities have a positive effect on cash holdings while liquid asset substitutes, size and leverage have a negative effect on cash. Besides they find that the higher the bank debt, the lower the cash holdings. This is an indication that more monitoring by banks or other creditors leads to lower cash holdings. Opler et al. (1999) examine determinants of cash holdings of public US firms and find that firm size is negatively related to cash holdings. Larger firms tend to hold lower levels of cash due to better access to capital markets and enhanced debt service capabilities. Opler et al. (1999) and Ozkan and Ozkan (2004) find a positive relationship between cash flows and cash ratios. Since cash flow data is scarce, operating income is used as proxy for cash flow. Operating income is defined as earnings before interest and taxes (EBIT). Ozkan and Ozkan (2004) find that firms with high operating income volatility have a higher probability of a cash shortage which induces managers to hold more cash. Dittmar et. al (2003) show that liquid assets serve as direct substitute for cash and cash equivalents suggesting a negative relationship between cash holdings and the level of other liquid assets.
The tables 1 and 2 provide an operationalization of the control variables for the corporate group and subsidiary regressions.
which is total debt divided by total assets. The market-to-book ratio is labeled MTB computed as the book value of assets minus the book value of equity, plus the market value of equity scaled by total assets. Operating income is defined as earnings before interest and taxes (EBIT) and labeled OP_INC. Operating income volatility labeled LOG_VOLAT is the natural logarithm of the standard deviation of EBIT which is defined as earnings before interest and taxes. It is measured over 3 years including the current year of analysis and scaled by the average total assets of this period. The total number of subsidiaries belonging to a corporate group may also affect its cash holdings due to relatively higher monitoring problems. I assume that control problems may be greater in corporate groups with a high number of subsidiaries compared to corporate groups with only a few subsidiaries. Therefore, I include the variable #SUBS.
IND_DUM represents the industry dummy variable. It takes on the value 1 if the corporate group is operating in the specified industry and 0 if otherwise. This measure is based upon two-digit US SIC core codes. With the exception of LOG_VOLAT and #SUBS, all remaining control variables are lagged by one year to avoid possible endogeneity problems.
4.3. Model description and justification
Following Opler et. al (1999), I estimate multiple regression models to examine the effects of the independent variables on both consolidated cash of the corporate group and unconsolidated cash of the subsidiary. The regression models are specified as follows:
(1) Corporate group-level regression
LOG_CASHi,t = α0+ β1EU_SUBSi + β2NON_EU_SUBSi + β3WO_SUBSi + β4MO_SUBSi
+ β5REL_SUBSi + β6DIV_SUBSi+ β7#SUBSi+ β8LOG_SIZEi,t-1 + β9LEVi,t-1 + β10MTBi,t-1
+ β11OP_INCi,t-1 + β12LOG_VOLATi,t+ εi,t
where the dependent variable LOG_CASHi,tdenotes the consolidated cash and cash equivalents of
the corporate group and α0is the constant. This is followed by the independent variables EU
subsidiaries, Non-EU subsidiaries, wholly-owned subsidiaries, majority-owned subsidiaries, related subsidiaries, diversified subsidiaries, total number of subsidiaries and the control variables size, leverage, market-to-book, operating income and operating income volatility. The error term is denoted by εi,t.
(2) Subsidiary-level regression
LOG_CASHi,t = α0+ β1EU_DUMi+ β2WO_DUMi+ β3MO_DUMi+ β4 REL_DUMi
+ β5DIV_DUMi+ β6LOG_SIZEi,t-1+ β7OP_INCi,t-1+ β8LOG_VOLATi,t+ β9LOG_LIQi,t+ εi,t
where the dependent variable LOG_CASH denotes the unconsolidated cash and cash equivalents of the subsidiary. On the right hand side of the equation α0is the constant which is followed by the
dummy variables EU, wholly-owned, majority-owned, relatedness and diversification and the control variables size, operating income, operating income volatility and liquidity. The error term is
represented by εi,t.
In both models the dependent variable LOG_CASH is measured in 2010 (year-end).When analysing cash ratios, endogeneity problems can occur since cash holdings can affect other variables such as leverage or growth opportunities. Considering the regression model, this means that the
Moreover, in multiple regression analysis, the residuals associated with the dependent variable can be heteroscedastic. Therefore, I estimate White heteroscedasticity adjusted standard errors (Hayes and Cai, 2007).
Finally, I include industry dummy variables in each regression model to account for differences in corporate cash holdings that are industry-specific. These dummy variables are based upon two digit US SIC core codes and take on the value 1 if the firm operates in the specified industry and 0 if otherwise. Two digit US SIC codes identify the major industry group.5The majority of corporate
5. Data description 5.1. Sample selection
The sample consists of 896 listed UK parent firms (see Table 3) and their 2.934 domestic and 1.013 EU-located subsidiaries (with complete financial data). I select public firms (corporate groups) because they have to follow IFRS rules for consolidation.6By taking only publicly listed firms, I can
assume that all firms use the same rules to consolidate their financial statements and can ensure consistency for the dependent variable LOG_CASH. Otherwise cash and cash equivalents would be differently reported under different accounting standards leading to a biased dependent variable in the corporate group regression model.
To retrieve the relevant data for the analysis, I use the Orbis database. Financial data (e.g. cash and cash equivalents or asset data, etc.) are collected for the end of 2010. Subsidiary ownership data, however, are not provided for specific dates or time periods. The Orbis database reports only the dates when changes in subsidiary ownership occured. Therefore, the information about subsidiary ownership refers to May 2012, when the data was retrieved.
When comparing older databases with newer databases, newer database versions show a higher number of recorded subsidiaries than older versions due to the fact that the company database has been extended and more companies have been added over time. As a result, I cannot infer if the number of subsidiaries has increased because the parent firm increased its investment in foreign subsidiaries or established new subsidiaries abroad or if this increase stems from new additions of subsidiaries by the database provider. In this respect, simply assuming ownership stability over time would be misleading. Therefore, a panel analysis would yield biased results and cannot be applied. Considering these data limitations a cross-section analysis is employed.
Data of parent firms and their subsidiaries are merged based on BvD ID Codes. The location of subsidiaries is identified based on ISO country codes7. These codes are also used to group
subsidiaries into specific regions (domestic, EU and Non-EU). Therefore, all firms with missing BvD ID codes are removed from the sample which is necessary to link parent firms and subsidiaries.
6Private UK firms have the option to choose between UK GAAP and IFRS rules concerning consolidation.
(The National Archives. Retrieved November 22, 2012 from: http://www.dti.gov.uk/bbf/financial-reporting/standards/page21457.html)
7Country Codes Standard - ISO 3166. Retrieved December 12th, 2012 from
I remove firms with unknown ownership since it is not evident to which firm they belong. Furthermore, all firms with unknown ownership percentages are excluded since it cannot be determined to which extent these firms are owned and therefore controlled by the parent firm. Moreover, firms with missing US SIC industry codes are removed since it cannot be identified if the parent firm and subsidiary operate in the same or a different industry.
Financial firms are excluded since they are legally obliged to meet standardized capital requirements set by the UK Financial Services Authority (FSA) and the Basel Accords. Moreover, they have high holdings of marketable securities which are included in cash and cash equivalents. Utilities are excluded as these firms are affected by regulatory requirements that do not apply to regular industrial firms.
As the literature mainly addresses corporate groups, I remove firms that have no subsidiaries at all and therefore cannot be classified as a corporate group. 298 listed firms from the initial sample (1.644) are removed since they have no subsidiaries.
6. Empirical results
This section provides the empirical results of the corporate group and subsidiary regressions. The descriptive statistics for the corporate group and subsidiary variables can be found in the tables 4 and 5, respectively. For the regression results, the tables 6 and 7 display the variables, the estimated coefficients and their corresponding p-values (in parentheses).
6.1. Results of the corporate group regressions
The first regression model tests six independent variables that measure control aspects as suggested by the literature. The variables EU_SUBS and NON_EU_SUBS are included to test whether control problems associated with multinationality lead to higher consolidated cash holdings of corporate groups. WO_SUBS and MO_SUBS are included to test whether a higher level of equity ownership in subsidiaries reduces monitoring problems leading to lower subsidiary cash holdings and thereby lower consolidated cash holdings of the corporate group. The variables REL_SUBS and DIV_SUBS are incorporated to test whether more difficulties to monitor and control unrelated or diversified subsidiaries lead to higher cash holdings of subsidiaries and eventually of the corporate group.
As outlined in Table 6, the first independent variable EU_SUBS is significant at the 1% level and therefore provides support for the first hypothesis H1a. The estimated coefficient is significant and
positive, indicating that the higher the number of EU subsidiaries belonging to a corporate group, the higher the consolidated cash of the corporate group. Due to a higher geographical distance, language barriers, higher communication and travelling costs, EU subsidiaries are more difficult to monitor by the UK parent firm compared to domestic subsidiaries.
Strong support is also found for hypothesis H1bwhich states that the higher the number of Non-EU
With the inclusion of the variables EU_SUBS and NON_EU_SUBS, I differentiate between domestic, EU and Non-EU subsidiaries, where high ratios indicate high multinational activity and low ratios high domestic activity. Domestic subsidiaries have a relatively lower geographical distance from
headquarters compared to EU and Non-EU subsidiaries. Hence, communication and travelling costs are relatively higher for corporate groups with high multinational activity. This in turn leads to less efficient monitoring and leaves possibilities for managerial entrenchment at EU and Non-EU subsidiaries. As a result, foreign subsidiary cash levels are built up resulting in higher consolidated cash holdings. The empirical results support this relationship between the degree of multinationality and consolidated cash holdings where higher domestic activity leads to lower consolidated cash and higher EU and Non-EU activity lead to higher consolidated cash holdings.
The results reported above are in line with Foley et al. (2007) and Ramirez and Tadesse (2009) who find that higher multinational activity leads to higher consolidated cash holdings. However, their explanation for higher cash ratios of multinationals is different since they do not put forward agency theory in explaining higher cash ratios of MNCs but better investment opportunities that
multinationals can exploit abroad. The authors of both papers argue that higher and also more volatile investment opportunities abroad induce subsidiary managers to hold more cash leading to higher consolidated cash holdings. Therefore, I include the variable MTB (market-to book ratio) and LOG_VOLAT (operating income volatility) in the regression model to control for higher and more volatile investment opportunities. After controlling for these effects, I still find that cash holdings of corporate groups with relatively higher multinational activity are still higher compared to corporate groups with relatively lower multinational activity.
withholding taxes on dividend flows within the corporation8. Since double taxation within the EU is
avoided, corporations with higher multinational activity in the EU are less inclined to leave cash in their subsidiaries and repatriate the foreign income instead. In this way excess cash levels in the subsidiaries are avoided leading to relatively lower consolidated cash holdings. In contrast, corporations with high Non-EU activity face higher tax costs on repatriating profits and therefore leave cash in their subsidiaries. According to Foley et al. (2007), these higher cash holdings in the subsidiaries are not offset by lower cash holdings of the parent firm leading to relatively higher consolidated cash holdings. The high statistical significance as well as the positive and large
coefficient of NON_EU_SUBS are in line with this explanation. These results propose that tax costs on repatriation are a strong determinant of consolidated cash holdings. However, when I control for repatriation taxes by analysing subsidiaries located in the EU only, in which these tax costs are irrelevant, I still find a positive relationship between the degree of multinationality and the level of cash holdings of the corporate group. Consequently, these findings show that beyond a tax cost explanation for higher cash holdings of MNCs, agency costs associated with international activity can be regarded as a relevant determinant.
Overall, the empirical results provide support both for an agency cost and for a tax cost explanation of higher consolidated cash holdings of MNCs.
Hypothesis H2aproposes that higher equity ownership and thereby higher control in subsidiaries
belonging to a corporate group results in lower consolidated cash of the corporate group. I argue that higher monitoring and control by the corporate headquarters limits agency problems in subsidiaries leading to lower consolidated cash levels. To test whether this hypothesis can be supported, I compute the variables WO_SUBS (wholly-owned subsidiaries) and MO_SUBS (majority-owned subsidiaries). As can be seen in Table 6, the coefficients of both variables have a negative sign indicating that higher ownership in subsidiaries results in a lower consolidated cash ratio. However, both variables are statistically insignificant, so I reject hypothesis H2a. Hence, the level of equity
ownership in subsidiaries is not found to be a determinant of consolidated cash holdings.
To test whether the ease of monitoring and control with which the parent firm can effectively govern subsidiary managers can be regarded as a determinant of consolidated cash holdings, hypotheses H3a
and H4aare formulated. As explained in section 3.3. and 3.4, specific subsidiary characteristics such as
8Council Directive 2003/123/EC of 22 December 2003 amending Directive 90/435/EEC on the common system
being either a related or unrelated subsidiary or a focused or diversified subsidiary may play a role in complicating monitoring by the parent firm and would allow subsidiary managers to build up cash levels.
I hypothesize in H3athat the higher the relative number of related subsidiaries belonging to a
corporate group, the lower the consolidated cash of the corporate group. The empirical findings reported in Table 6 do not support this hypothesis. Relatively lower information asymmetries and information costs between parent firms and related subsidiaries do not seem to result in lower cash holdings in the parent firm. Likewise, relatively higher information asymmetries do not allow managers of unrelated subsidiaries to build up cash levels.
Hypothesis H4astates that the higher the number of diversified subsidiaries belonging to a corporate
group, the higher the consolidated cash of the corporate group. Existing theory proposes that the difficulty for headquarters to acquire, consolidate and interpret information of diversified
subsidiaries may be higher compared to focused subsidiaries (Buckley and Strange, 2010). Besides, it may be relatively more difficult for headquarters management to adequately determine an optimal cash level for a diversified subsidiary. This leaves opportunistic subsidiary managers with
opportunities to accumulate cash.
The findings of the regression analysis reject this hypothesis (H4a) since the variable DIV_SUBS is
statistically insignificant. Hence, subsidiary diversification does not have an effect on cash holdings. A possible explanation maybe found from the fact that diversified firms may hold less cash compared to focused firms since the risk of a cash shortage is lower (Duchin, 2010). However, even this effect is not confirmed by the findings, which suggests that a relationship between subsidiary diversification and consolidated cash holdings cannot be supported.
The p-value of OP_INC is insignificant. Therefore, a positive relationship between cash flows and cash ratios, as proposed by Opler et al. (1999), cannot be confirmed. Besides, firms with high operating income volatility have a higher probability of cash shortage which induces managers to hold more cash (Ozkan and Ozkan, 2004), which is confirmed by a positive sign of LOG_VOLAT.
6.2. Results of the subsidiary regressions
The second regression model focuses on control and subsidiary cash holdings. As there is not much data available for Non-EU subsidiaries, I focus on EU subsidiaries only. I include five independent variables that relate to control. These variables are used to test whether agency costs of cash are present at the subsidiary level. The dependent variable represents unconsolidated cash of the subsidiary. I compute the dummy variable EU_DUM to test whether control problems associated with multinationality lead to higher unconsolidated cash holdings. The dummy variables WO_DUM and MO_DUM are included to test whether a higher level of equity ownership by the parent reduces subsidiary cash holdings. Finally, the dummy variables REL_DUM and DIV_DUM are incorporated to test whether a higher difficulty to monitor and control subsidiaries leads to higher subsidiary cash holdings.
Table 7 displays the results of the subsidiary regressions. EU_DUM denotes whether a subsidiary is domestic or located in one of the (remaining) 26 EU countries. The findings suggest that compared to domestic subsidiaries, EU subsidiaries do not have relatively higher cash holdings as is hypothesized in H1c. The argument that control problems associated with internationalization such as language
barriers, communication costs and travelling costs lead to higher subsidiary cash holdings is not supported by these empirical findings. From this, I infer that parent firms are equally effective in governing EU and domestic subsidiaries. The headquarters management seems to be able to judge the cash requirements and determine cash levels of EU subsidiaries adequately. As a result, the information asymmetries between parent firms and EU subsidiaries may be comparable to those between parent firms and domestic subsidiaries.
between headquarters managers of MNCs and external shareholders exist which lead to higher MNC but not higher subsidiary cash holdings.
Hypothesis H2bstates that a higher degree of equity ownership by the parent firm leads to lower
subsidiary cash holdings. It is argued that headquarters management limits the discretion of subsidiary managers to build up cash when it has high controlling stakes (measured by the equity ownership level). As displayed in Table 7, this hypothesis is confirmed by the regression results. Both variables WO_DUM and MO_DUM have large and negative coefficients and are significant at the 1% level. In comparison to partially-owned subsidiaries, the parent firm possesses substantial controlling stakes in majority-owned and full stakes in wholly-owned subsidiaries. The findings suggest that controlling parent firms make use of their voting power to reduce subsidiary cash holdings.
The last two hypotheses H3band H4bare formulated to analyse if specific subsidiary characteristics
complicate monitoring and control by the parent firm.
Hypothesis H3bproposes that cash holdings of related subsidiaries are lower compared to cash
holdings of unrelated subsidiaries. The regression results provide support for this hypothesis since the p-value of the variable REL_DUM is significant around the 5% level and the coefficient has a negative sign. This indicates that headquarters managers are able to effectively control related subsidiaries but not unrelated subsidiaries. Headquarters may not have the same relevant industry information and expertise which an unrelated subsidiary has. Besides, headquarters managers may have only limited access to specific industry information and may have difficulties to interpret the information correctly. Therefore, it may be difficult for headquarters managers to determine optimal cash levels of unrelated subsidiaries leading to relatively higher cash holdings.
Hypothesis H4bstates that cash holdings of diversified subsidiaries are higher compared to cash
As displayed in Table 7, the firm-specific control variable LOG_SIZE is significant at the 1% level. LOG_SIZE has a negative effect on subsidiary cash holdings as proposed by Ferreira and Vilela (2004). The p-value of OP_INC is insignificant and the coefficient has a negative sign which does not support the view that a positive relationship between operating income and cash ratios exists. The coefficient of LOG_VOLAT is positive and significant around the 5% level. This indicates that firms with relatively higher operating income volatility hold more cash (Ozkan and Ozkan, 2004). LOG_LIQ has a small and negative coefficient indicating that liquid assets serve as direct substitute for cash and cash
equivalents (Dittmar et. al, 2003), however the p-value is insignificant.
6.3. Robustness tests
Table 8 outlines a robustness test for the corporate group regression. I re-estimate the regression model with the newly computed variables WO_MO_SUBS, REL_SUBS_2 and DIV_SUBS_2. Where WO_MO_SUBS is the number of subsidiaries in which the parent owns >50% -100% of equity divided by all subsidiaries. REL_SUBS_2 is measured with US SIC primary codes at 3 digit level and
DIV_SUBS_2 is measured with US SIC primary codes at 4 digit level. The results in the regression table remain basically unchanged. The sign and the size of the coefficients as well as the p-values are in line with those in the first regression model (Table 6).
7. Conclusion
This thesis investigates monitoring and control problems and their effects on cash holdings using a sample of 896 listed UK parent firms and their 2.934 domestic and 1.013 EU-located subsidiaries.
I study control within corporate groups on three different dimensions. First, I examine monitoring and control problems associated with multinational activity and their effect on cash holdings. Second, I analyse control problems associated with limited equity ownership in subsidiaries and their effect on cash holdings. Finally, I test whether specific subsidiary characteristics such as being either unrelated or diversified complicate monitoring by the parent firm and how this affects cash holdings.
I hypothesize that control problems within corporate groups lead to managerial entrenchment as it leaves opportunities for subsidiary managers to build up cash levels that exceed an optimum level. Using multiple regression analysis, I first test the relationship between consolidated cash of the corporate group and independent variables that relate to control problems as suggested by existing literature. Second, I examine this relationship for the unconsolidated cash of the subsidiary.
The empirical findings of this thesis provide support for a tax cost motive and the agency theory of internationalization in explaining higher consolidated cash holdings of MNCs. In line with Foley et al. (2007), tax costs on repatriating foreign income have a significant positive effect on cash holdings of corporate groups with high multinational activity outside the EU. In addition, I control for
repatriation taxes by analysing cash holdings of subsidiaries located within the EU only, in which these tax costs are irrelevant. I still find a positive relationship between the degree of
multinationality and the level of cash holdings of the corporate group. Moreover, I find that especially a greater geographical distance between headquarters and foreign subsidiary (EU versus Non-EU subsidiaries) results in higher consolidated cash holdings of the corporate group.
Considering the results of the subsidiary regressions, I find a negative relationship between equity ownership of the parent firm and subsidiary cash holdings indicating that higher controlling stakes in a subsidiary prevent subsidiary managers from building up cash.
With respect to multinationality however, I find that compared to domestic subsidiaries, EU
subsidiaries do not have higher cash holdings. Thus, the hypothesis that control problems associated with internationalization such as language barriers, geographical distance and associated
communication and travelling costs lead to higher subsidiary cash holdings is not supported by the findings of this study.
In addition, I find no empirical support that subsidiary diversification and relatedness can be regarded as relevant determinants of subsidiary cash holdings.
7.1. Suggestions for future research
For subsidiaries outside the European Union only little data are available. Especially financial data for subsidiaries outside the EU are scarce, as the Orbis database includes primarily company data of European firms. Hence, further research could compare cash holdings of Non-EU subsidiaries with those of domestic and EU- subsidiaries to examine which effects (tax costs, agency costs, or investment opportunities) can be regarded as relevant determinants of their cash holdings.
8. References
Andersson U., and Forsgren M., 1996, Subsidiary embeddedness and control in the multinational corporation, International Business Review 5(5), 487–508.
Bates, T.W., Kahle, K.M. and Stulz, R.M., 2009, Why Do U.S. Firms Hold So Much More Cash than They Used To?, Journal of Finance 64, 1985-2011.
Billett, M., Mauer, D., 2003, Cross-Subsidies, External Financing Constraints, and the Contribution of the Internal Capital Market to Firm Value, Review of Financial Studies 16, 1167–201.
Brav, O., 2009, Access to capital, capital structure, and the funding of the firm, Journal of Finance 64, 263–308.
Buckley, P. J. and Strange, R., 2010, The governance of the multinational enterprise: insights from internalization theory, Journal of Management Studies, 460-470.
Chang, E., and Taylor, S. M., 1999, Control in multinational corporations (MNCs): The case of Korean manufacturing subsidiaries, Journal of Management 25(4), 541-565.
Dischinger, M. and N. Riedel, 2009, There’s No Place like Home: The Profitability Gap between Headquarters and their Foreign Subsidiaries, CESifo Working Papers, No. 2866, University of Munich.
Dittmar, A., Mahrt-Smith J., and Servaes, H., 2003, International Corporate Governance and Corporate Cash Holdings, Journal of Financial and Quantitative Analysis 38, 111-133.
Dittmar, A., and Mahrt-Smith J., 2007, Corporate governance and the value of cash holdings, Journal of Financial Economics 83, 599-634.
Doukas, J.A., 1995, Overinvestment, tobin's Q and gains from foreign acquisitions, Journal of Banking and Finance 19(7), 1285-1303.
Duchin, R., 2010, Cash Holdings and Corporate Diversification, Journal of Finance 65 (3), 955–992.
Ferreira, M.A. and Vilela, A., 2004, Why Do Firms Hold Cash? Evidence from EMU Countries, European Financial Management 10, 295-319.
Filatotchev, I., and Wright, M., 2011, Agency Perspectives on Corporate Governance of Multinational Enterprises, Journal of Management Studies 48, 471–486.
Foley, C.F., Hartzell, J.C., Titman, S., Twite, G., 2007, Why do firms hold so much cash? A tax based explanation, Journal of Financial Economics 86, 579-607.
Gaur, A.S., and Lu, J.W., 2007, Ownership strategies and survival of foreign subsidiaries: impacts of institutional distance and experience, Journal of Management 33, 84–110.
Gertner, R., D. Scharfstein, and J. Stein, 1994, Internal Versus External Capital Markets. Quarterly Journal of Economics 109, 1211–30.
Ghoshal, S., and Bartlett, C., 1990, The multinational corporation as an interorganizational network, Academy of Management Review 15 (4), 603-625.
Hayes, A. F., and Cai, L., 2007, Using heteroscedasticity-consistent standard error estimators in OLS regression: An introduction and software implementation, Behavior Research Methods 39, 709-722.
Harford, J., Mansi, S.M., and Maxwell, W.F., 2008, Corporate governance and firm cash holdings in the US, Journal of Financial Economics 87, 535-555.
Jensen, M.C., 1986, Agency costs of free cash flow, corporate finance and takeovers, American Economic Review 76, 323– 329.
Jiraporn, P., Kim, Y. S., Davidson, W. N. and Singh, M., 2006, Corporate governance, shareholder rights and firm diversification: an empirical analysis, Journal of Banking and Finance 30, 947–63.
Kalcheva, I. and K. Lins, 2007, International Evidence on Cash Holdings and Expected Managerial Agency Problems, Review of Financial Studies 20, 1087-1112.
Lee, I. and Wyatt, S., 1990, The effects of international joint ventures on shareholder wealth, The Financial Review 25, 641–9.
Mikkelson, W. H., and Partch, M., 2003, Do persistent large cash reserves hinder performance? , Journal of Financial and Quantitative Analysis 38, 257-294.
Modigliani, F., and Miller, M. , 1958, The Cost of Capital, Corporation Finance and the Theory of Investment, American Economic Review 48 (3), 261–297.
Myers, S. , and Majluf, N., 1984, Corporate financing and investment decisions when firms have information that investors do not have, Journal of Financial Economics, 187–221.
Nohria, N. and Ghoshal, S., 1994, Differentiated fit and shared values, Strategic Management Journal 15, 491-502.
O’Donnell, S.W., 2000, Managing foreign subsidiaries: agents of headquarters, or an interdependent network?, Strategic Management Journal 21, 525–548.
Opler, T., L. Pinkowitz, R. Stulz, and R. Williamson, 1999, The determinants and implications of corporate cash holdings, Journal of Financial Economics 52, 3-46.
Prahalad, C. K., and Doz, Y. L., 1987, The multinational mission: Balancing local demands and global vision, Journal of International Business Studies 19, 304-306.
Ramírez, A., and Tadesse, S., 2009, Corporate cash holdings, uncertainty avoidance, and the multinationality of firms, International Business Review 18(4), 387-403.
Roth, K. and A. J. Morrison, 1990, An empirical analysis of the integration-responsiveness framework in global industries, Journal of International Business Studies 22 (4), 541-61.
Stulz, R. M., 1990, Managerial discretion and optimal financing policies, Journal of Financial Economics 26, 3-27.
Stulz, R., and W. Wasserfallen, 1995, Foreign equity investment restrictions, capital flight, and shareholder wealth maximization: Theory and evidence, Review of Financial Studies 8, 1019-1057.
Tong, Z., 2008, Firm Diversification and the Value of Corporate Cash Holdings, Journal of Corporate Finance 17 (3), 741-758.
Von Eije, J. H., 2012, What Causes Differences in Cash Holdings between Listed and Unlisted Firms Around the World?, Working papers series, Available at SSRN: http://ssrn.com/abstract=2010053
Wade, M.R., and Gravill, I.J., 2003, Diversification and performance of Japanese IT subsidiaries: a resource-based view, Information and Management 40(4), 305–316.
9. Tables
Table 1: Variable operationalization for the corporate group regressions
Variable Measurement
Dependent LOG_CASH Consolidated cash Natural logarithm of cash and cash equivalents scaled by net assets (which are defined as total assets minus cash and cash equivalents)
Independent EU_SUBS Multinational activity within EU (27)
Number of EU (27) subsidiaries/ All subsidiaries
Independent NON_EU_SUBS Multinational activity outside the EU
Number of NON-EU subsidiaries/ All subsidiaries
Independent WO_SUBS Full equity ownership
Number of wholly owned subsidiaries/ All subsidiaries
Independent MO_SUBS Majority equity ownership
Number of majority owned subsidiaries/ All subsidiaries
Independent WO_MO_SUBS >50% -100% equity ownership
Number of subsidiaries in which the parent owns >50% -100% of equity / All
subsidiaries
Independent REL_SUBS Relatedness Number of related subsidiaries / Sum of related and unrelated subsidiaries. Measured with US SIC primary codes at 2 digit level
Independent REL_SUBS_2 Relatedness Number of related subsidiaries / Sum of related and unrelated subsidiaries. Measured with US SIC primary codes at 3 digit level
Independent DIV_SUBS Degree of diversification
Independent DIV_SUBS_2 Degree of diversification
Number of diversified subsidiaries/ All subsidiaries. Measured with US SIC primary codes at 4 digit level
Independent #SUBS Number of
subsidiaries
Total number of subsidiaries belonging to a corporate group
Control LOG_SIZE Firm size Natural logarithm of the book value of total assets lagged by one year
Control LEV Leverage Total debt divided by total assets lagged by one year
Control MTB Market-to-book
ratio
Book value of assets minus the book value of equity, plus the market value of equity scaled by total assets and lagged by one year.
Control OP_INC Operating income Earnings before interest and taxes (EBIT) scaled by total assets lagged by one year Control LOG_VOLAT Operating Income
Volatility
Natural logarithm of the standard deviation of EBIT which is defined as earnings before interest, taxes and depreciation. It is measured over 3 years including the current year and scaled by the average total assets of this period.
