Excess cash holdings and investment: the moderating
roles of financial constraints and managerial
entrenchment
Her-Jiun Sheu
a, Shiou-Ying Lee
b,caDepartment of Banking and Finance, National Chi Nan University, Nantou County, Taiwan
bInstitute of Business and Management, National Chiao Tung University, Taipei, Taiwan
cDepartment of Marketing and Logistics Management, Taipei Chengshih University of Science
and Technology, Taipei, Taiwan
Abstract
Our study investigates the relationship between excess cash holdings and invest-ment behaviour under two dimensions of financial constraints and managerial entrenchment, based upon a sample of Taiwanese firms operating in an environ-ment characterized by poor legal protection for investors, with data covering the years 2000–2006. We find that excess cash is significantly correlated with capital expenditure, particularly for firms financially constrained and with severe mana-gerial entrenchment. However, the evidence shows that excess cash is insensitive to R&D expenditure under these two dimensions.
Key words: Excess cash; Financial constraints; Managerial entrenchment; Overinvestment; Underinvestment
JEL Classification: G31, G34
doi: 10.1111/j.1467-629X.2012.00500.x
1. Introduction
When studying corporate finance, it is important to see how the determinants complement each other to influence a firm’s investment decisions. Various forces prevent a firm from pursuing its optimal investment level when the presumption of a perfect market is violated. Information asymmetries and agency problems are
The authors would like to thank Chih-Liang Liu, Fu-Ru Yang, Tzu-pu Chang and semi-nar participants at 2010 the 18th Annual Conference on PBFEAM and 2011 European Financial Management Association Annual Meetings in Braga for helpful comments and suggestions.
the most important factors influencing investment efficiency (Stein, 2003), whereas cash holdings are strongly related to a firm’s investment when facing these fric-tions. On the one hand, the adverse selection problem arises, because managers are reluctant to issue undervalued securities because of information asymmetries, which leads to underinvestment. Cash holdings also help firms with high external financing costs (i.e. financially constrained firms) take up positive net present value (NPV) projects. Such a phenomenon therefore makes an investment sensitive to cash holdings. On the other hand, empire-building preferences lead to overinvest-ment, causing entrenched managers to spend all available funds on investment (Jensen, 1986). This also leads to higher investment because of cash holdings.
Prior studies have shown how cash holdings are associated with investment when considering either financial constraints or corporate governance. Almeida
et al.(2004) indicate that financially constrained firms tend to save cash, whereas
unconstrained firms do not. Consistent with the costly external finance view of Faulkender and Wang (2006) and Pinkowitz and Willionson (2004) in that cash holdings are more valuable for constrained than for unconstrained firms, Denis and Sibilkov (2010) explain that higher cash holdings allow financially con-strained firms to undertake value-increasing projects that might be bypassed. Another strand of research addresses the agency cost hypotheses, as recent stud-ies document that poor corporate governance is detrimental to the value of cor-porate cash holdings (Dittmar et al., 2003; Pinkowitz et al., 2006; Dittmar and Mahrt-Smith, 2007). Harford et al. (2008) propose that self-interested managers of US firms act to avoid the attention of activists, and therefore, they choose to spend cash quickly on capital expenditures and acquisitions.
La Porta et al. (1998) argue that greater investor protection increases inves-tors’ willingness to provide funds in exchange for securities, hence reflecting in a lower cost for the availability of external financing. Khurana et al. (2006) docu-ment that financial constraints are more pronounced in less financially developed countries with weaker institutions that protect investors. In addition, Harford
et al. (2008) propose true entrenchment requires low legal shareholder rights.
Thus, it is much easier for managers in a weaker legal protection environment to act contrary to the interests of shareholders.
McLean et al. (2012) indicate that investor protection affects firm-level resource allocation. They provide evidence by analysing a large sample of firms from 44 countries during the period 1990–2007 and showing that investment sen-sitivity to q and external finance relative to q are stronger in countries with greater investor protections, because high q firms can easily obtain external finance to fund investments, while investment sensitivity to cash flow is higher in countries with fewer investor protections. Their findings highlight that strong investor protection laws predict accurate share prices, reduce financial con-straints and encourage efficient investment. According to their country-level estimation, the investment-q sensitivity is 0.127 in the United States compared with 0.058 for Taiwan, while the investment-cash flow sensitivity is 0.077 in the United States compared with 0.686 for Taiwan.
The low investment-q and high investment-cash flow sensitivity of Taiwan1 thus provide an interesting setting for us to consider, compared with the United States, how differences in a country-level investor protection environment inter-acting with a firm-level corporate governance structure make liquidity decisions different. Weaker investor protection laws and a less-developed financial market in Taiwan make firms rely even more on internal resources, and we argue the effect is stronger when facing financing constraints. In addition to cash flow, cash
is a good candidate to fund investment.2Besides, the true entrenchment effect is
more likely to influence managers’ liquidity decisions in countries with poor investor protection. Therefore, we argue that high cash holdings could more quickly facilitate overinvestment at the expense of shareholders in poorly gov-erned firms in Taiwan. Furthermore, another question arises: do highly entrenched managers still engage in overinvestment when their firms face higher external financing costs?
Using a study sample covering the years 2000–2006, we focus on excess cash used for two types of investment expenditure: (i) capital expenditure; and (ii) research and development (R&D) expenditure. We apply six criteria constraints and develop a managerial entrenchment index to capture the governance mecha-nism in Taiwan.
Splitting our sample for testing the costly external finance hypothesis in accor-dance with the financial constraint criteria, our results show that capital expendi-tures have statistically significant sensitivity to excess cash, which is stronger for constrained firms, providing support for the underinvestment argument. How-ever, although excess cash is found to be significantly and positively correlated with R&D expenditure for both constrained and unconstrained firms, we are unable to provide any consistent results to suggest that the relationship is stron-ger for constrained firms.
Splitting our sample for testing the agency problem hypothesis in accordance with the managerial entrenchment index, our results find that the sensitivity of investment to excess cash has a positive sign under higher entrenchment, thereby indicating a tendency towards overinvestment. When we apply low (high) institu-tional blockholdings to proxy for an inefficient (efficient) corporate governance mechanism, the results also show that capital expenditure-excess cash sensitivity is stronger for low blockholding firms. This confirms the overinvestment argu-ment.
Our empirical evidence contributes to the extant literature on investment behaviour when using excess cash by simultaneously accounting for both
1McLean et al. (2012) use five different measures of investor protection and two variables to investigate how easily firms can issue equity. Taiwan displays lower investor protection (for example, the investor protection index is 0.547 in Taiwan versus 1.000 in the United States) and less easy access to equity markets than the United States.
2
underinvestment arising from financing constraints and overinvestment arising from managerial entrenchment. The evidence in this study reveals that both dimensions affect investment-excess cash sensitivity. If a firm suffers from financ-ing constraints, then excess cash can finance investment projects that the firm may have previously given up as a result of a shortage of internal resources. Fur-thermore, if the firm is characterized by managerial entrenchment, then excess cash may induce such managers to invest in projects that could prove detrimen-tal to shareholder wealth. When the ownership share held by institutional inves-tors is used as a proxy for the quality of a firm’s corporate governance structure, the two dimensions become even more significant, thereby indicating both prob-lems of underinvestment and overinvestment actually coexist on investment-excess cash sensitivity.
Our results indicate that in contrast to capital expenditure, R&D expendi-ture is insensitive to excess cash under these two dimensions. Our findings suggest that financially constrained firms do not invest in large R&D expen-diture when holding excess cash, while entrenched managers are less likely to
overinvest in R&D, essentially as a result of their risk aversion.3 Owing
to the high adjustment costs of R&D, Taiwanese firms use cash reserves to finance R&D expenditures regardless of their financial status. We provide evi-dence that cash holdings are positively associated with real R&D investment spending for both financially constrained and unconstrained firms. We explain that firms prefer to use internal funds on R&D investment instead of expen-sive equity sources. The findings in Taiwan are different from Brown and Petersen’s (2011) results that US firms most likely facing financing frictions rely extensively on cash holdings to smooth R&D, whereas less financially constrained firms do not, because equity finance appears to be the principal source of funds.
The rest of this paper proceeds as follows. Section 2. Theoretical background and hypotheses describes the theoretical background and develops the hypothe-ses. Section 3. Data and variable construction describes data and variables that are used in the study. Section 4. Methodology details the research methodology. Section 5. Empirical results presents our empirical findings. Section 6. Conclu-sions concludes the paper.
2. Theoretical background and hypothesis
Myers and Majluf (1984) suggest that, in an imperfect market, information asymmetry between corporate insiders and outsiders invariably results in costly external financing, which causes the adverse selection problem. Managers may be forced to give up positive NPV projects, because they are not willing to raise
3
It should be noted that compared with capital expenditure on property, plant and equip-ment, R&D expenditure is typically viewed as high-risk investment (Kothari et al., 2002).
external capital by issuing underpriced securities. Therefore, cash flow and cash can benefit those firms facing external financing constraints by funding necessary expenditures, which makes their investment sensitive to the availability of
inter-nal funds (Stein, 2003; Franzoni, 2009).4
Faulkender and Wang (2006) argue that liquidity provides a benefit for constrained firms, thereby demonstrating that the marginal value of cash holdings is more valuable for financially constrained firms than for uncon-strained ones. Denis and Sibilkov (2010) interpret prior findings to mean that constrained firms with higher cash holdings are more likely to use cash to increase investment in positive NPV projects and that marginal investment is more valuable to financial constrained than for unconstrained firms. Brown and Petersen (2011) provide evidence that firms facing financing frictions tend to rely heavily on cash holdings to smooth their R&D spending, essentially because cash provides a buffer to R&D from financial shocks and avoids the high adjustment costs of R&D. This reasoning results in our first hypothesis as follows.
H1: After controlling for investment opportunities and cash flow, the sensitivity of
corporate investment expenditure to excess cash is more positive for financially con-strained firms than for unconcon-strained firms.
Managers with empire-building preferences will use all available resources on investment projects beyond a level that would maximize shareholder value (Jensen, 1986). As noted by Myers and Rajan (1998), when managers have power over corporate decisions and are not constrained by legal provisions or effective external monitoring, it is much easier for cash reserves to be expropri-ated. Indeed, even when insiders cannot expropriate directly, they may use cash to finance negative NPV projects for their personal benefit – that is, they have a
tendency for overinvestment based upon empire-building.5
Cross-border studies provide evidence to show that weak shareholder rights are associated with higher cash holdings (Dittmar et al., 2003; Pinkowitz and
4
Fazzari et al. (1988) argue that cash flow is the primary capital for financially con-strained firms and that the sensitivity of cash flow to investment is stronger for constrained than for unconstrained firms. Kaplan and Zingales (1997) show that invest-ment-cash flow sensitivity is not necessarily higher for firms that are more constrained because Tobin’s q is a noisy proxy for marginal q. If cash flow contains more information about future investment opportunities, then even less constrained firms are likely to adjust their investment in response to the informativeness of cash-flow shocks and exhibit higher investment-cash flow sensitivities (Alti, 2003). Rather than focusing on investment-cash flow sensitivity, Almeida et al. (2004) find that constrained firms display a significantly positive cash-flow sensitivity of cash, while unconstrained firms do not.
5
Fresard and Salva (2010) explain that this occurs when insiders do not have sufficient power to expropriate outsiders or when legal protections effectively constrain such expro-priation.
Willionson, 2004), with the value of cash holdings lower in those countries (Pinkowitz et al., 2006; Kalcheva and Lins, 2007).
Dittmar and Mahrt-Smith (2007) and Harford et al. (2008) provide evi-dence to show how corporate governance influences the decision on the ways in which cash should be spent. Dittmar and Mahrt-Smith (2007) note that firms with poor governance structures and higher excess cash experience lower operating performance because of a rapid dissipation of cash, implying that, under the conditions of serious agency problems, excess cash reduces the pressure on managers to operate efficiently. Harford et al. (2008) show that, relative to their industry peers, poorly governed firms with higher levels of excess cash tend to increase their capital and acquisition expenditures, while reducing R&D investment. This reasoning results in our second hypoth-esis as follows.
H2: After controlling for investment opportunities and cash flow, the sensitivity of
corporate investment expenditure to excess cash is more positive for higher entrenchment firms than for lower entrenchment firms.
Stein (2003) argues that financial slack is important to investment, but it is far from clear as to whether this relationship is attributable to financing constraints or empire-building. With regard to the sensitivity of investment to cash flows, although the costly external financing and agency conflict theories are essentially equivalent, their policy implications differ markedly. Therefore, the two dimen-sions may well coexist in a unified model that considers both underinvestment and overinvestment (Stein, 2003; Franzoni, 2009).
With the purpose to find that the sensitivity of investment to financial slack depends on either under- or overinvestment, Franzoni (2009) nests both the financial constraints and empire-building models into one specification. The results demonstrate that a reduction in liquidity leads financially constrained firms to underinvestment, which has a negative effect on shareholder value. Conversely, when managers pursue their personal interests, the reduction in internal resources, which is less costly for outside investors, has a positive effect on firm value. The evidence shows that underinvestment is more rele-vant for the entire sample. Analogous to Franzoni (2009) and Xu et al. (2012) demonstrate that listed family firms in China are prone to underinvest-ment, as opposed to overinvestunderinvest-ment, and that political connectedness could reduce the level of investment-cash flow sensitivity for those firms with finan-cial constraints, thereby providing further support for the underinvestment argument.
Acknowledging that underinvestment and overinvestment may coexist within the same firm (Stein, 2003; Franzoni, 2009), we take the financing constraints variable and the managerial entrenchment variable into account and try to find out whether the two dimensions coexist to influence the relationship between excess cash and investment expenditure.
3. Data and variable construction 3.1. Data
Our sample includes all nonfinancial listed firms in Taiwan, covering the years from 2000 to 2006. After discarding all observations with incomplete data, we are left with a total sample of 2596 firm-year observations for subsequent analy-sis. Corporate governance data and other company information are collected from the Taiwan Economic Journal (TEJ) database.
We follow Opler et al. (1999) approach to measure the normal level of cash holdings. Excess cash is the difference between actual and predicted normal cash
holdings.6 Those firms with excess cash greater than zero are adopted as our
sample to test our hypotheses.
3.2. Managerial entrenchment measures
3.2.1. Construction of the managerial entrenchment index
Managerial entrenchment has gained considerable attention as a result of its implications for corporate governance. Managers entrench themselves by pursu-ing self-interest policies that do not maximize shareholder value (Shleifer and Vishny, 1989). We adopt the following proxies, each of which has predicted asso-ciations with managerial entrenchment.
1. Affiliated board seats (Aff_Bd): Board seats are classified as being affiliated when they are held by the firm’s largest shareholder, by the identifiable rela-tives of the largest shareholder or by legal representarela-tives from other
6
We use natural log of cash to net assets as the dependent variable (Ln (Cash/NA)), while the independent variables include natural log of assets (Size), cash flow to net assets (CashFlow/NA), net working capital to net assets (NWC/NA), the mean industry standard deviation in cash flow over assets over the previous 5-year period (Indus-trySigma), market value to net assets (MV/NA), R&D to net assets (RD/NA), total debt to net assets (Leverage/NA), capital expenditure to net assets (Capex/NA) and a dummy indicating whether the firm paid dividends in that year (Dividend) and includes industry and year indicators to estimate the normal level of cash holdings. To avoid the problem of endogeneity, we follow the procedure of Dittmar and Mahrt-Smith (2007) to employ 3-year lagged sales growth (SalesG) as an instrument variable for MV and find that SalesGis a good proxy for investment opportunity. Our regression results shows that Ln (Cash/NA) = )0.251 + 0.549CashFlow/NA)0.284Size)0.365 NWC/NA+1.245Indus-trySigma+ 1.746RD/ NA+ 0.465MV/NA+ 0.105Capex/NA+0.912Leverage/NA + 0.143 Dividend. We find that smaller firms and firms with larger cash flows, growth opportunities, R&D expenditure and leverage tend to hold more cash, as do dividend paying firms and firms with lower net working capital.
companies controlled by the largest shareholder (Yeh and Woidtke, 2005).
Aff_Bd is defined as the number of affiliated directors divided by the total
number of directors.
2. Independent directors (Ind_Dir): Independent directors have expertise in man-agement and decision-making and are less subject to agency conflicts (Fama and Jensen, 1983). Ind_Dir is measured as the number of independent board seats divided by the total number of board seats.
3. Separation of ownership and control (Sep_OC): La Porta et al. (1999) suggest the separation of ownership and control can benefit controlling shareholders to control a firm’s operations with a small direct stake in cash-flow rights.
Sep_OCis equal to 1 if the voting rights of the controlling shareholders are
higher than cash-flow rights; otherwise, 0.7
4. Cash compensation ratio (CCR): Berger et al. (1997) argue that CEOs with higher levels of cash compensation are more likely to be entrenched and will therefore seek to avoid risk. Listed firms in Taiwan pay stock bonuses as incentives for employees. Therefore, we define CCR as the proportion of cash
salary to total compensation paid to CEOs.8
5. CEO_duality: On the basis of the agency cost hypothesis, Jensen (1993) points out that CEO duality may hinder board effectiveness, while also increasing agency costs. Nevertheless, the ‘stewardship theory’ suggests that CEO duality may benefit firm value, because it provides a unity of leadership structure (Donaldson, 1990). The CEO_duality dummy variable is equal to 1 if the CEO is also the chairman of the board; otherwise, 0.
We use principal component analysis (PCA) to construct a managerial
entrenchment index.9 As reported by Florackis and Ozkan (2009), PCA
helps to control for problems of multicollinearity that may arise when several governance and control variables are incorporated within the empirical models. PCA automatically produces the weights so that the
7
Claessens et al. (2000) find that almost 80 per cent of firms in Taiwan have managers and directors who are controlling shareholders. Moreover, Yeh and Woidtke (2005) indi-cate that Taiwan is characterized by having a high level of ownership concentrated in the largest controlling shareholders, as well as a significant divergence in control and owner-ship. It is pervasive for the controlling shareholders of Taiwan firms to utilize dominant control power to exploit minority shareholders.
8Total compensation comprises cash salary plus the value of stock bonuses.
9Callahan et al. (2003) construct an index of management involvement in director nomi-nations using PCA and ten governance variables, while Florackis and Ozkan (2009) also utilize the same approach, combining governance variables to construct a corporate gov-ernance measure in UK firms.
measure will capture the largest proportion of variance in the underlying
data.10
3.2.2. Institutional blockholdings
We follow prior studies to use the blockholdings of institutional investors as an additional measure of the quality of corporate governance (Dittmar and
Mahrt-Smith, 2007; Franzoni, 2009).11 Institutional blockholders are
better-informed investors and have strong incentives to devote resources to monitoring, because they have large amounts of investment at stake (Shleifer and Vishny, 1986). Therefore, the presence of institutional investors creates effective monitor-ing and hence mitigates agency conflicts. Higher (lower) institutional blockhold-ings are associated with better (worse) monitoring and discipline on management and thus indicate better (poor) governance structure.
3.3. Financial constraints criteria
We select six approaches associated with firm-level financial status as proxies for financial constraints below.
1. Firm size (Size): Almeida et al. (2004) state that small firms have difficulties in raising capital within the market, because they are less well known. We clas-sify financially constrained firms if their book value of total assets is below the median level in the year.
2. Dividend dummy (Dividend): Compared with constrained firms, unconstrained firms are more likely to have higher payout ratios (Almeida et al., 2004). We classify financially constrained firms if they did not pay cash dividends in the
year.12
3. Cash flow (CashFlow): Firms with larger internal cash flows may find it easier to obtain external financing, because such firms are invariably perceived by lenders as being less risky (Leland and Pyle, 1977). We classify financially constrained firms if their cash flow normalized by the start-of-year book assets is below the sample median level (Babenko et al., 2011) in the year. 10
Taking a combination of the above five governance variables based upon PCA, with the selection of the first principal component, the managerial entrenchment index = 0.613* Aff_Bd-0.626*Ind_Dir+0.208*Sep_OC+0.398*CCR)0.177*CEO_duality. The negative weight of CEO_duality provides support for the stewardship theory (Donaldson, 1990), which argues that leadership unity that effectively reduces entrenchment is beneficial to firm performance.
11Institutional blockholdings are defined as equity ownership by an institutional share-holder with ownership greater than 5 per cent.
12
We find the empirical results are unchanged if we use the average payout ratio as a con-straint criteria.
4. Firm age (Age): As suggested by Hadlock and Pierce (2010), firm age and firm size are the two variables with the greatest relevance to financial
con-straints.13We estimate the ages of Taiwanese firms since their listing on the
TSE or OTC market, classifying financially constrained firms if their ages are below the sample median level in the year.
5. Bank loans (Loan): Shen and Wang (2005) investigate that firms in Taiwan
are less financially constrained when they have a strong bank relationship.14
We use bank loans as a proxy for bank relationships and categorize finan-cially constrained firms if their total bank loans are below the sample mean for each year.
6. KZ index (KZ): Following Almeida et al. (2004), we divide our sample into
two groups according to the measure (which is called the KZ index)15
developed by Kaplan and Zingales (1997). We classify financially con-strained firms if their KZ index ranking is above the sample medium level in the year.
Table 1 provides descriptive statistics. Panel A of Table 1 shows summary sta-tistics of the variables used to predict the normal level of cash holdings and the firm-specific variables to proxy for financial constraints. All ratios are winsorized at 1 per cent and 99 per cent to reduce the impact of outliers. Panel B reports summary statistics of the variables we use to construct the managerial entrench-ment index and the institutional blockholdings.
Table 2 reports the Pearson correlation coefficients for the proxies of financial constraints and managerial entrenchment, showing that the
correla-tion coefficients on the financial constraints proxy range from )0.654 to
0.593, each with statistical significance. Although the high correlations imply that the measures are picking up similar information, it appears that each measure picks up certain unique information (Denis and Sibilkov, 2010). Finally, blockholdings are uncorrelated with the managerial entrenchment index.
13
Hadlock and Pierce (2010) use qualitative information on firms to develop an index of financial constraints. After evaluating several common sorting variables, they clude that firm size and age (SA index) appear to be closely related to financial con-straints.
14
Shen and Wang (2005) evaluate the bank relationship using three proxies: the number of banks that a firm engages with for its borrowing, the loan amounts and the loan dura-tion. The results remain robust regardless of which of these is used as the proxy for bank relationship.
15The KZ index, which is defined normalizing variables by total assets in the previous period, is calculated as )1.002*CashFlow + 0.283*Q + 3.139*LTD-39.368*Dividends )1.315*Cash, where CashFlow is the ratio of cash flow to total assets, Q is Tobin’s q, LTDis ratio of long-term debt to total assets, Dividends is the ratio of total dividends to book assets, Cash is the ratio of cash to total assets.
Table 1
Descriptive statistics
Variables N Mean Median SD 25% 75%
Panel A XCash/Assets 3776 0.101 0.065 0.109 0.028 0.134 Size 3776 15.183 15.027 1.151 14.284 15.860 CashFlow/Assets 3776 0.074 0.070 0.080 0.032 0.117 NWC/Assets 3776 0.142 0.132 0.171 0.019 0.259 IndustrySigma 3776 0.059 0.048 0.044 0.029 0.075 RD/Assets 3776 0.020 0.009 0.029 0.000 0.026 MV/Assets 3776 1.279 1.121 0.543 0.914 1.480 SalesG 3776 0.219 0.143 0.389 )0.002 0.350 Capex/Assets 3776 0.059 0.041 0.059 0.016 0.084 Leverage/Assets 3776 0.458 0.462 0.168 0.342 0.578 Dividend 3776 0.614 1.000 0.487 0.000 1.000 Age 3776 23.009 22.000 11.012 14.000 30.000 Loan (millions) 3107 1152.437 445.555 2631.944 150.000 1079.532 KZ index 3776 )0.405 )0.128 1.022 )0.879 0.312 Panel B Aff_Bd 2597 0.683 0.667 0.207 0.571 0.833 Ind_Dir 2597 0.092 0.000 0.146 0.000 0.200 Sep_OC 2597 0.772 1.000 0.420 1.000 1.000 CCR 2597 0.719 0.899 0.326 0.417 1.000 CEO_duality 2597 0.324 0.000 0.468 0.000 1.000 Managerial entrenchment index 2597 )0.127 0.048 1.275 )1.004 0.908 Blockholdings 3512 0.334 0.301 0.209 0.160 0.475
Panel A reports summary statistics of the variables used to predict the normal level of cash holdings and the firm-specific variables to proxy for financial constraints based on a sample covering the 2000–2006 period. XCash/Assets is the ratio of excess cash to assets; Size is natural log of assets; CashFlow/Assetsis the ratio of cash flow to assets; NWC/Assets is the ratio of current assets exclud-ing cash minus current liabilities to assets; IndustrySigma is the mean industry standard deviation in cash flow over assets over the previous 5-year period; RD/Assets is the ratio of R&D to assets; MV/ Assetsis the ratio of the market value of equity plus the book value of assets minus the book value of equity divided by book value of total assets; SalesG is the growth in sales over the previous 3-year period; Capex/Assets is the ratio of capital expenditure to assets; and Leverage/Assets is the ratio of the sum of long-term debt and current liabilities to assets. Dividend is a dummy variable which is equal to 1 if the firm paid a common dividend in that year; otherwise 0. Age is number of years since the firm is listed on the TSE or OTC market; Loan is the total amount of bank loans; KZ index is a linear combination of financial variables developed by Kaplan and Zingales (1997); all ratios are winsorized at the 1% and 99% levels. Panel B reports summary statistics of the variables we use to construct the managerial entrenchment index: Aff_Bd is the number of affiliated directors divided by the total number of directors; Ind_Dir is the ratio of the number of independent board seats divided by the total number of board seats; Sep_OC is a dummy which takes the value of 1 if the voting rights of the controlling shareholders exceeds cash-flow rights; otherwise 0.; CCR is cash compensa-tion ratio; CEO_duality is a dummy which is equal to 1 if the CEO is also the chairman of the board; otherwise 0. The managerial entrenchment index is calculated as the weighted sum of the above five components. Blockholdings is equity ownership of institutional blockholders.
Table 2 Correlat io n m atrix o f finan cial constrain ts a nd man age rial en tr enc h ment Variables Size Dividen d Ca sh Flo w Ag e K Z L oa n M -ind ex Blo ckh olding s Size 1 Di vi de n d 0.0 5 1 (0 .000 ) 1 CashFlow 0.5 9 3 (< 0 .00 0 ) 0 .0 53 (< 0.00 0) 1 Age 0.2 5 5 (< 0 .00 0 ) 0 .0 09 (0 .036 ) 0 .0 36 (0 .01 7 ) 1 KZ ) 0 .24 1 (<0 .0 00 ) ) 0 .65 4 (< 0 .0 00 ) ) 0 .16 5 (< 0 .0 00 ) ) 0 .1 9 4 (0.00 0) 1 Loan 0 .44 3 (<0 .0 00 ) ) 0.0 6 3 (< 0 .00 0 ) 0 .1 66 (< 0.00 0) 0 .1 6 5 (<0.00 0 ) ) 0. 12 0 (<0 .0 00 ) 1 M-i n d ex ) 0.1 3 1 (< 0 .00 0 ) 0 .0 20 (< 0.00 0) ) 0 .08 3 (< 0 .0 00 ) ) 0. 13 0 (<0 .0 00 ) 0 .0 4 3 (0 .0 00 ) ) 0. 04 6 (<0 .0 10 ) 1 Blo ckh ol di ng s 0 .21 5 (<0 .0 00 ) ) 0.1 1 0 (< 0 .00 0 ) 0 .1 77 (< 0.00 0) ) 0 .0 3 2 (0.03 5) 0 .0 6 5 (<0.00 0 ) 0 .1 11 (<0.00 0) ) 0 .0 3 5 (0.02 0) 1 This table reports the co rrelations between th e v ar iable s in o u r a nalysis o f the fin ancial const raint s a nd manage rial en tr enc h ment, w ith the data co v ering the 2000– 2 006 p eriod h aving be en o b tained from the TEJ. W e exclude co mpanies w ithin the finan cial industries. See section F inancial const raints criteri ao ft h e text for full d efinitions of the financial const raints criteria: Si ze ; Divide nd ; CashFlow ; Age; KZ an d Loan . M-in d ex refers to the m an agerial entrenchme n t index; and Bloc kh ol din g s is equity ownersh ip o f institu ti onal b lo ckholders. T he p -values (r epo rted in pare ntheses) ar e b as ed o n rob u st st anda rd err o rs .
4. Methodology
We consider two measures of firm investment: (i) capital expenditure; and (ii)
R&D expenditure.16We further note that R&D has a number of characteristics
that differ from ordinary investment, as detailed below.
First, R&D investment is particularly subject to financing constraints, because it is firm-specific and difficult to evaluate (Himmelberg and Petersen, 1994). Sec-ond, R&D involves substantial adjustment costs, essentially because most costs of R&D are related to wage payments to highly skilled workers. Firms facing financial friction should therefore manage their liquidity to maintain smooth R&D (Brown and Petersen, 2011).
We test our first hypothesis on the investment-excess cash sensitivity with the division of our sample into financially constrained and unconstrained firms being undertaken in accordance with our six financial constraints. We test our second hypothesis on the investment-excess cash sensitivity with the division of our sam-ple into higher entrenchment and lower entrenchment firms being undertaken
according to our managerial entrenchment index.17
The equation is as follows:
Iit¼ b0þ b1XCashi;t1þ b2Qi;t1þ b3CashFlowi;tþ xi;t c
þ Year Dummies þ Industry Fixed Effects þ ei;t ð1Þ
Here, Iitrepresents the investment expenditure by firm i in year t; and XCashi,t)1
is the start-of-year excess cash, scaled by the start-of-year book assets. The latter is measured by lagging the data by a year, which reduces the extent of the poten-tial problem of endogeneity arising from simultaneous determination of these
variables. Term Qi,t)1 is the start-of-year market-to-book ratio to control for
investment opportunity, which is the market value of equity plus the book value of assets minus the book value of equity divided by the book value of total
assets. CashFlowi,t is cash flow normalized by the start-of-year book assets.
Lastly, xi,tis a set of control variables representing the firm’s financial status.
To test the costly external finance hypothesis, we expect that, ceteris paribus, the coefficient on XCash is more positive for financially constrained than for unconstrained firms. To test the empire-building hypothesis, we expect that, cete-ris paribus, the coefficient on XCash is more positive for higher entrenchment than for lower entrenchment firms.
According to Stein (2003), we demonstrate the logic behind taking into account both the costly external finance model and the empire-building model. Our 16When there are missing values on R&D, it is set as being equal to zero.
17
We also classify low (high) blockholding firms if their institutional blockholdings are below (above) the sample medium level to represent the firms with a weaker (better) gov-ernance structure.
financial constraints and managerial entrenchment proxies are included in Eqn (2), along with their interactions with excess cash. The equation is as follows:
Iit¼ b0þ b1XCashi;t1þ b2FCi;t1þ b3XCashi;t1 FCi;t1þ b4Mi;t1
þ b5XCashi;t1 Mi;t1þ b6Qi;t1þ b7CashFlowi;tþ xi;tc
þ Year Dummies þ Industry Fixed Effects þ ei;t ð2Þ
Here, FCi,t –1is an indicator variable equal to 1 for firms with greater financial
constraints at the start of year, and Mi,t –1is an indicator variable equal to 1 for
firms with higher managerial entrenchment at the start of year. All other vari-ables are the same as defined in Eqn (1).
When considering both external financing cost and agency cost on the relation-ship between excess cash and investment, we include the interactions of
XCash*FCand XCash*M in the model. If the coefficient on XCash*FC is
signifi-cantly greater than zero, then it reflects that excess cash could be used to increase investment for constrained firms (underinvestment). If the coefficient on
XCash*Mis significantly greater than zero, then it reflects that entrenched
man-agers spend more on investments out of the excess cash (overinvestment). When both XCash*FC and XCash*M are statistically greater than zero, it reflects that excess cash could be used to avoid underinvestment for constrained firms, whereas excess cash could also be spent towards overinvestment if constrained firms have a poor governance structure.
Numerous empirical studies focus on the ways in which financial status affects
investment.18We rely upon these studies on investment decision-making for the
control variables representing financial status. They include growth in sales
(SalesG), firm size (Size) and total leverage (Leverage).19
The investment regression includes both industry and year fixed effects. The industry fixed effects may help to minimize the likelihood of excess cash affecting
different types of investment friction across different industries.20The year fixed
18
See, for example, Fazzari et al. (1988), Kaplan and Zingales (1997) and Denis and Sibilkov (2010).
19
Owing to the editorial issues, we do not report all the control variables on our tables. Detailed results are available from the authors upon request.
20
Denis and Sibilkov (2010) suggest that industry fixed effects rather than firm fixed effects are more appropriate for testing the hypothesis that greater cash holdings allow constrained firms to invest more than unconstrained firms could, because constrained firms rationally choose to maintain higher cash levels to fund investment. They do not think a firm’s year-to-year variation in cash reserves is associated with investment. Fur-thermore, if firms with higher cash holdings invest more, then a firm’s fixed effect specifi-cation will not pick up the effect, because it would disregard the higher level of investment as being a fixed effect. On the basis of the above reasons, we adopt industry fixed effects instead of firm fixed effects.
effects are used to control for the macroeconomic effects that could conceivably affect investment decision-making.
5. Empirical results
Table 3 reports the fixed effects regressions of investment on excess cash under financial constraints criteria. Panel A of Table 3 reports the sensitivity of capital expenditure to excess cash. As compared to unconstrained firms, under five of the six financial constraints, excess cash is found to have a signif-icantly positive correlation with the capital expenditure for constrained firms. The coefficients on XCash indicate that about NT$0.036 (p = 0.064) to NT$0.079 (p = 0.007) of each one NT dollar of excess cash is allocated to increase capital expenditure for constrained firms compared with -NT$0.078 (p = 0.000) to -NT$0.047 (p = 0.017) for unconstrained firms. These results are largely in agreement with the first hypothesis, in that the dependence of excess cash on capital expenditure is stronger for constrained than for uncon-strained firms.
Panel B of Table 3 presents the sensitivity of R&D expenditure to excess cash, from which we find that for both constrained and unconstrained firms, excess cash has a positive and significant correlation with R&D expenditure. The coefficients on XCash indicate that about NT$0.033 (p = 0.000) to NT$0.054 (p = 0.000) of each one NT dollar of excess cash is allocated to increase R&D expenditure for constrained firms compared with NT$0.020 (p = 0.000) to NT$0.049 (p = 0.001) for unconstrained firms. However, only two of the six constraints criteria show that the constrained firms would spend more excess cash on R&D spending. The results in Panel B indicate that excess cash enables both constrained and unconstrained firms to increase R&D
investment.21
Table 4 reports the fixed effects regressions of investment on excess cash under managerial entrenchment. Panel A of Table 4 shows the investment-excess cash sensitivity when we use managerial entrenchment index as the measure for cor-porate governance. The dependence of XCash on capital expenditure is positive
0.022 (p = 0.223) for higher entrenchment firms and )0.023 (p = 0.302) for
lower entrenchment firms, providing the empire-building preference of highly entrenched managers to use excess cash on capital expenditure although the dif-ference is not significant. The R&D-excess cash sensitivity results show that NT$0.044 (p = 0.000) of each one NT dollar of excess cash is allocated to increase R&D expenditure for higher entrenchment firms compared with NT$0.031 (p = 0.000) for lower entrenchment firms. However, we are unable to
21
We also restrict the sample to the firms with R&D expenditure greater than zero to test all of our empirical settings and find the results are unchanged.
Table 3 Fixe d effects re g ressions of in vestment on ex ce ss cash un der financ ial constraints crite ria Variables Si ze Di vi de n d C a sh Fl ow Ag e L o a n K Z FC UN FC U N FC UN FC U N FC UN FC UN Pa ne l A : C a p ex XC ash 0. 036 (0 .0 6 4 )* ) 0. 047 (0 .0 17) ** 0. 062 (0 .0 21) ** ) 0. 020 (0 .209 ) 0. 0 5 3 (0. 001 )** * ) 0. 078 (0 .0 00)* * * 0. 007 (0 .7 18) 0. 0 0 6 (0 .744 ) 0. 0 5 4 (0. 013 )** ) 0. 075 (0 .1 0 2 ) 0. 07 9 (0 .0 07) *** ) 0. 02 2 (0 .1 75) Q 0. 018 (0 .0 0 0 )* * 0. 009 (0 .0 26) ** 0. 033 (0 .0 00) *** 0. 0 0 7 (0 .061 )* 0. 0 2 1 (0. 000 ) * ** 0. 008 (0 .0 54)* 0. 012 (0 .0 05) *** 0. 0 1 6 (0 .000 )** * 0. 0 0 9 (0. 045 )** 0. 021 (0 .0 1 3 )* * 0. 02 4 (0 .0 00) *** 0. 009 (0 .0 15) ** Ca sh F lo w 0. 194 (0 .0 0 0 )* ** 0. 329 (0 .0 00) *** 0. 187 (0 .0 00) *** 0. 2 8 7 (0 .000 )** * 0. 1 1 9 (0. 000 )** * 0. 390 (0 .0 00)* * * 0. 248 (0 .0 00) *** 0. 2 7 4 (0 .000 )** * 0. 2 7 3 (0. 000 )** * 0. 303 (0 .0 0 0 )* ** 0. 24 9 (0 .0 00) *** 0. 255 (0 .0 00) *** N o . o f O bs. 119 3 1 4 0 3 8 24 1 772 117 8 1 4 1 7 1 219 1 302 125 3 4 4 2 11 48 1 448 Ad j. R 2 0. 213 0. 347 0 .2 6 0 0 .2 80 0. 1 9 8 0 .309 0 .2 7 7 0 .2 83 0. 2 7 2 0 .377 0. 34 5 0 .2 2 9 p -val ue di ffe re n ce in co effi ci en ts o n XC ash 0. 021 ** 0 .0 0 8** * 0 .0 00* ** 0 .7 2 4 0 .0 27* * 0 .0 0 0 *** Pa ne l B : R & D XC ash 0. 042 (0 .0 0 0 )* ** 0. 031 (0 .0 00) *** 0. 037 (0 .0 00) *** 0. 0 3 9 (0 .000 )** * 0. 0 3 3 (0. 000 )** * 0. 038 (0 .0 00)* * * 0. 052 (0 .0 00) *** 0. 0 2 0 (0 .000 )** * 0. 0 5 4 (0. 000 )** * 0. 049 (0 .0 0 1 )* ** 0. 03 3 (0 .0 01) *** 0. 038 (0 .0 00) *** Q 0. 014 (0 .0 0 0 )* ** 0. 007 (0 .0 00) *** 0. 012 (0 .0 00) *** 0. 0 1 1 (0 .000 )** * 0. 0 1 5 (0. 000 )** * 0. 008 (0 .0 00)* * * 0. 015 (0 .0 00) *** 0. 0 0 6 (0 .000 )** * 0. 0 1 1 (0. 000 )** * 0. 013 (0 .0 0 0 )* ** 0. 00 8 (0 .0 00) *** 0. 012 (0 .0 00) *** Ca sh F lo w 0. 007 (0 .4 3 6 ) 0. 016 (0 .0 42) ** ) 0. 02 3 (0 .0 34) ** 0. 0 1 9 (0 .016 )** ) 0. 0 0 9 (0. 352 ) 0. 021 * (0 .0 57) 0. 005 (0 .6 03) 0. 0 1 8 (0 .013 )** 0. 0 0 5 (0. 587 ) 0. 005 (0 .7 4 1 ) 0. 00 7 (0 .3 88) 0. 008 (0 .3 78) N o . o f O bs. 119 3 1 4 0 3 8 24 1 772 117 8 1 4 1 7 1 219 1 302 135 3 4 4 2 11 48 1 448 Ad j. R 2 0. 377 0. 395 0 .3 7 1 0 .3 78 0. 3 7 5 0 .375 0 .3 7 7 0 .3 23 0. 3 5 7 0 .413 0. 34 6 0 .3 8 1 p -val ue di ffe re n ce in co effi ci en ts o n XC ash 0. 006 *** 0 .1 6 2 0 .9 79 0 .0 0 0** * 0 .7 74 0. 63 7 Th is ta bl e re p o rt s the co effi ci ents o n the fix ed eff ec ts of in ve st m en t on ex ce ss ca sh un de r the fina nc ia l co n st rai n ts cr it er ia , p re se n ti n g th e m o de ls se p a rat el y fo r the gro u p s of fi n an ci al cons tra in ed (F C ) a n d u n co n st ra in ed (U N) fi rm s. T h e d ep en d en t v a ria b le in P a n el A is ca p it a l ex p en d it u re (Ca pe x ), an d the de pe nde nt va ri abl e in Pa ne l B is R & D ex p en di tur e (R& D ); ea ch of th e d ep en d en t v a ri a b le s is n o rm a li ze d b y th e st a rt -o f-y ea r b o o k a ss et s. XC a sh re fe rs to th e b eg inni ng -of -ye ar ex ce ss ca sh hol di ngs o f a fir m n o rm a li ze d b y th e st art -of -ye a r b ook a ss et s. T h e co ntr o l v ari a bl es ar e a s fol lo w s: Q is th e b eg in n in g -o f-y ea r m a rk et -to-boo k rati o ; Ca sh F lo w is ca sh flow no rm a li ze d by th e star t-o f-y ea r b o o k a ss et s; Si ze is n a tu ra l log o f a sse ts ; Sa le sG is th eg ro w thi ns a le s o v er th e p re v iou s thr ee -ye a r p er iod, a n d Le ve rag e is th e su m of lo ng-te rm d eb t a n d cu rr en t li a bi li ti es n o rm al iz ed by th e sta rt -of -ye ar bo ok as se ts. T he p -val ue s b as ed on robu st st and a rd er ro rs a re re p o rte d in p a re n the se s. * **, ** , a nd * in d ic a te stat is ti ca l si gni fi ca nc e a t the 1, 5, 10% le ve ls , re spe ct iv el y . A ll es ti m a ti o ns in cl u d e in d us try a nd ye ar in d ic a to rs as we ll as in te rc ep t te rm .
Table 4 Fixed effe cts regressions of inve stmen t on ex cess cash under m anagerial entren chment Va ri ab le s Panel A Panel B Cape x R &D Ca pex R&D Highe r entrench ment Lower entrenchme nt Hig h er en tren chment Lower entrenc h ment Low blockholding Hi gh b loc kholding Low blockholding Hig h b loc kholding XC a sh 0.02 2 (0 .223) ) 0.023 (0 .3 0 2 ) 0 .044 (0.0 0 0 )* * * 0.03 1 (0 .000) * * * 0 .0 3 7 (0 .031 )* * ) 0.06 1 (0 .006) * * * 0 .0 4 0 (0 .000 )* * * 0.036 (0.0 0 0 )* * * Q 0.01 1 (0 .002) * * * 0 .0 1 5 (0.008 )* * * 0.011 (0.0 0 0 )* * * 0.01 0 (0 .000) * * * 0 .0 2 1 (0 .000 )* * * 0.006 (0.21 9 ) 0 .0 1 1 (0 .000 )* * * 0.011 (0.0 0 0 )* * * Ca sh F lo w 0.26 2 (0 .000) * * * 0 .2 5 4 (0.000 )* * * 0.009 (0.2 1 1 ) 0 .01 4 (0 .243) 0 .1 7 6 (0 .000 )* * * 0.347 (0.00 0 )* * * 0 .0 1 5 (0 .064 )* 0.005 (0.6 0 3 ) No . o f O bs . 1 7 2 4 8 7 2 1724 872 1 3 90 1205 1 3 90 1205 Ad j. R 2 0.29 8 0 .2 4 5 0.390 0.35 9 0 .2 6 8 0.317 0 .3 8 3 0 .371 p -v al ue d iff er enc e in co efficien ts o n XC a sh 0.22 7 0 .317 0 .0 0 2** * 0 .4 6 9 P a ne l A re po rt s the co effi ci en ts o n th e fi xe d eff ec ts o f inv es tm en t o n ex ce ss ca sh u n d er m a na g er ia l en tr en ch m ent , p re se n ti n g th e m o de ls se p a ra te ly fo r g rou ps o f h ig h er /l ow er en tr en ch -m ent fi rm s, w he re h ig h er (l ow er ) en tr en chm en t fi rm s a re d efin ed a s th o se w it h m a n a g er ia l en tr en ch m en t in d ex a b ov e (be lo w ) th e q ua rt il e. P a n el B re p or ts th e coe ffi ci en ts o n the fix ed effec ts o f in v estmen t o n excess cash u n d er in stitu tio n a l b lo ck h o ld in g s, p re se n tin g th e mo d els se p a rately fo r g ro u p s o f h igh /lo w b lo ck h o ld in g fi rms , w h er e h ig h (l o w ) bl o ck h o ldi ng fir m s ar e d efi ne d a s tho se wi th in st it ut io na l b lo ck h o ld in gs a b ov e (b el o w) th e m ed iu m . T h e d ep en d en t v a ri a b le s a re ca p ita l ex p en d itu re no rm al iz ed by th e st ar t-o f-y ea r b o o k a ss et s (C a pe x ); R & D ex pe nd it u re n or m a li ze d b y th e st a rt -o f-y ea r bo ok a ss et s (R & D ). XC a sh re fe rs to th e b eg in n ing -o f-ye ar ex ce ss ca sh h o ld in gs o f a fi rm no rm a li ze d by th e sta rt -o f-y ear b o o k a ss ets . Th e con tr o l v a ri a bl es ar e a s fol lo w s: Q is th e b eg in ni n g -o f-y ea r m a rk et -t o-b o o k ra ti o; Ca sh F lo w is ca sh flo w n o rm a li ze d by th e st a rt -o f-y ea r bo ok a ss et s; Size is n a tu ra l log o f a ss ets ; SalesG is th e g ro wt h in sa le s ov er th e p re v iou s th re e-y ea r p er io d , a n d Le verage is th e su m o f lo n g -t er m d eb t a n d cu rr en t li a bi li ti es n o rm a li ze d b y th e st a rt -o f-y ea r b o o k as se ts . T he p -v al ue s b as ed on ro b u st st a n d a rd er ro rs ar e re p o rt ed in p ar en th es es . * * * , * * a nd * in d ic a te st a ti st ic a l si g n ifi ca n ce a t th e 1 , 5 an d 1 0% le v el s, re sp ec tively. A ll es ti m a tio n s in clu d e in d u st ry a n d y ea r ind ic a tor s a s w el l a s in te rc ept te rm .
find evidence that R&D-excess cash sensitivity is stronger for higher
entrench-ment than for lower entrenchentrench-ment firms.22
Panel B of Table 4 presents the results of investment-excess cash sensitivity when we use institutional blockholdings as a robustness check. For low block-holding firms, NT$0.037 (p = 0.031) of each one NT dollar of excess cash is allocated to capital expenditure compared with -NT$0.061 (p = 0.006) for high blockholding firms. Thus, the findings confirm the entrenchment effect on capital expenditure-excess cash sensitivity. However, although excess cash is significantly positive to R&D for both high and low blockholding firms, we still cannot find that the sensitivity is higher for low blockholding than for high blockholding firms.
Table 5 presents the regression results examining the effect of excess cash on investment under financial constraints and managerial entrenchment. Panel A of Table 5 reports the sensitivity of capital expenditure to excess cash under the two dimensions. The coefficients on the interactions XCash*FC and
XCash*M indicate that (an) additional NT$0.072 (p = 0.007) to NT$0.124
(p = 0.014) of each one NT dollar of excess cash is used in capital expendi-ture for constrained than for unconstrained firms. On the other hand, (an) additional NT$0.049 (p = 0.063) to NT$0.073 (p = 0.046) of each one NT dollar of excess cash might be overinvested on capital expenditure when con-strained firms are poorly governed. We therefore conclude that firms that are small in size have lower cash-flow levels and have bank loans below the mean level tend to use excess cash to fund their capital expenditure. Moreover, when their managers are more highly entrenched, the firms will also have a tendency for overinvestment.
Panel B of Table 5 reports R&D-excess cash sensitivity under the two dimen-sions. The coefficients on XCash*FC are positively significant only for the size and age criteria. However, the coefficients on XCash*M are all found to be posi-tive, although only the size and loan criteria have statistical significance. Overall, there is no consistent evidence to show which dimension affects R&D-excess cash sensitivity.
Table 6 presents the regression results examining the effect of excess cash on investment under financial constraints and institutional blockholdings. Panel A
22
We also examine the cash holdings by adding the managerial entrenchment index in the predicted normal cash holding regression. The regression shows Ln (Cash/NA) = )1.742 + 2.361CashFlow/NA-0.097Size+0.327 WC/NA-0.972IndustrySigma+ 4.859RD/ NA+0.093MV/NA+0.099Capex/NA+0.603Leverage/NA+0.219Dividend-0.059M-inex. This means that highly entrenched firms in Taiwan hold lower cash reserves, which is con-sistent with the finding of Harford et al. (2008) in that managers in United States firms with a weaker governance structure spend cash quickly. We explain weak governance firms have lower cash reserves because managers spend cash on capital expenditures (over-investment).
Table 5 Fixed eff ects regre ssio ns of investme nt on excess cash unde r fi n a ncial co n stra ints and m anagerial entre nchme n t Vari abl es Size Dividen d CashFlow Age L o a n K Z Coeff. p -value C o eff. p -val ue Coeff. p -val u e Coeff. p -value Coeff . p -v a lu e C o eff . p -val ue Pa ne l A : C ap ex XCash ) 0.05 8* * 0 .0 19 ) 0.0 2 7 0 .1 66 ) 0. 08 1* ** 0 .0 0 1 ) 0.00 6 0 .80 4 ) 0.10 7* * 0 .0 45 -0 .0 28 0 .152 FC ) 0.00 1 0 .8 10 ) 0 .01 4* * * 0. 00 1 ) 0. 00 9* * 0 .0 32 0 .0 0 5 0 .2 21 0 .0 0 5 0 .3 3 5 ) 0. 00 6 0 .1 1 5 FC*XCash 0.07 2* ** 0.0 0 7 0 .0 81 ** 0 .0 1 3 0 .1 14 ** * 0 .00 0 ) 0.09 0 0 .72 7 0.12 4* * 0 .0 14 0 .1 1 0 * * * 0 .001 M ) 0.00 5 0 .2 89 ) 0.0 0 2 0 .6 13 ) 0. 00 5 0 .2 54 ) 0.00 3 0 .47 9 ) 0.00 9* 0 .0 9 1 ) 0. 00 3 0 .5 6 1 M*XCash 0.04 9* 0.0 6 3 0 .0 22 0 .4 0 4 0 .0 51 * 0 .05 4 0.03 1 0 .23 6 0.07 3* * 0 .0 46 0 .0 1 5 0 .572 N o . o f O b s. 2 5 9 62 5 9 62 5 9 52 5 2 1 1 7 9 52 5 9 6 Adj . R 2 0.28 1 0 .2 82 0 .2 8 4 0 .28 2 0.29 5 0 .2 83 Panel B : R &D XCash 0.01 5 0 .1 09 0.0 3 5 *** 0 .0 0 0 0 .0 32 ** * 0 .00 0 0.00 1 0 .95 0 0.02 5 0 .1 72 0 .0 3 6 * * * 0 .000 FC 0.00 1 0 .4 48 0.0 0 1 0 .3 50 ) 0. 00 2 0 .1 93 ) 0.00 1 0 .39 6 ) 0.00 3 0 .1 19 0 .0 0 1 0 .394 FC*XCash 0.02 9* ** 0.0 0 3 ) 0.0 1 8 0 .1 23 0 .0 0 1 0 .89 9 0.05 8* ** 0.00 0 0 .00 4 0 .8 1 4 ) 0. 02 2* 0.0 5 2 M ) 0.00 3* 0.0 9 5 ) 0.0 0 2 0 .1 50 ) 0. 00 2 0 .1 83 ) 0.00 2 0 .17 8 ) 0.00 4* * 0 .0 19 ) 0. 00 2 0 .1 5 1 M*XCash 0.01 7* 0.0 7 5 0 .0 12 0 .2 1 4 0 .0 10 0.29 7 0 .01 1 0.23 8 0 .04 6 * * * 0 .0 00 0 .0 1 3 0 .182 N o . o f O b s. 2 5 9 62 5 9 62 5 9 52 5 2 1 1 7 9 52 5 9 6 Adj . R 2 0.37 4 0 .3 71 0 .3 7 1 0 .38 7 0.37 6 0 .3 71 This ta ble repo rts the coeffic ie nts o n the fixe d effe cts of inve stmen t on ex ce ss cash unde r both fi n a ncial constrain ts a nd managerial en tr enc h ment. T he de pe n -dent variable in P a n el A is capital ex pen d iture (Capex ), while the d epe ndent variable in Panel B is R&D expendi ture (R&D ); all o f the dep endent variables are normalized by the start-of-year book assets. XCash re fers to the b eginning-o f-year excess cash hol dings o f a firm n ormalize d b y the st a rt-of-year book as se ts. FC is a d ummy variable for financial constrai nts w hich is equal to 1 (0) if the firm is fina ncially con strained (un constraine d); M is a dum m y va ri able whic h is equal to 1 (0 ) if the m a nagerial entren ch ment index is a bove (be low) the sample q uart ile. The control v ariable s are a s follows: Cash F low is ca sh flow normalized by the start-of-year boo k a ssets; Q is th e b egi nning-of-year m arket-to -bo o k rati o ; Si ze is natural log of asse ts; Sa le sG is th eg ro w thi ns a le s o v er the p re vious three-year pe riod, and Leverag e is the sum of long-te rm debt and curre nt liabilities normalized by the start-of-year b o o k a ssets. The p -values based o n robust standard er rors are reported in p are n theses. * **, * * a nd * ind icate statistical signifi ca nce a t th e 1, 5 a nd 10% levels, respective ly. A ll es ti m a -tio n s include ind u stry and y ea r in d ic ators a s w ell a s interc ept te rm. D etailed results are a vailable from the authors u p o n requests.
of Table 6 provides results on capital expenditure under the two dimensions.23 The coefficients on the interactions XCash*FC and XCash*Block indicate that (an) additional NT$0.044 (p = 0.099) to NT$0.113 (p = 0.000) of each one NT dollar of excess cash is used on capital expenditure for constrained than for unconstrained firms. Conversely, (an) additional NT$0.065 (p = 0.012) to NT$0.081 (p = 0.002) of each one NT dollar of excess cash might be overinvest-ed on capital expenditure for constrainoverinvest-ed firms with low institutional blockhold-ings.
Panel B of Table 6 shows the results on R&D under the two dimensions. The coefficients on the interaction XCash*FC are found to be significantly positive for two of the six criteria, whereas none of the other coefficients on the interac-tion XCash*Block are found to have any statistical significance. We still cannot find a consistent effect on R&D-excess cash sensitivity.
Compared with the results of Table 5, the significance of the interaction
XCash*Block in the results of Table 6 highlights the entrenchment effect on
investment-excess cash sensitivity, even though the firms are financially con-strained. The excess cash can help these firms take on profitable projects, but a part of it is still overinvested when constrained firms are less monitored by insti-tutional blockholders.
6. Conclusions
In this study, we aim to determine how excess cash affects firm investment expenditure in an imperfect market. Two hypotheses are developed to examine whether underinvestment arising from information asymmetry exists in finan-cially constrained firms and whether overinvestment arising from agency prob-lems exists in poor governance firms. Furthermore, considering both the costly external finance and empire-building dimensions, we attempt to find out whether these two dimensions coexist to affect investment-excess cash sensitivity.
Using six financial constraints approaches, we find that the dependence of excess cash for capital expenditure is stronger for constrained than for uncon-strained firms. This is mostly consistent with the argument of underinvestment. Nevertheless, excess cash is associated with R&D for both constrained and unconstrained firms.
We develop a managerial entrenchment index and find empire-building prefer-ences in capital expenditure for highly entrenched managers. The evidence holds for low blockholding firms when we use institutional blckholdings as a robust check. Nevertheless, R&D expenditure is unrelated to the problem of overinvest-ment.
23
When we use institutional blockholdings as another measure of corporate governance, the Block dummy is an indicator variable equal to 1 for firms less monitored by institu-tional blockholders at the start of year.
Table 6 Fixed eff ects regre ssio ns of investme nt on excess cash unde r financial co n stra ints and institutional b loc kholdings Va ri a b le s Si ze Di vi de n d C a sh Fl ow Ag e L oa n K Z Coeff. p -v a lu e C o eff . p -val u e C o eff. p -value Coeff. p -val ue Coeff . p -value Coeff. p -value Pa ne l A : C ap ex XCash ) 0. 05 7* ** 0. 00 9 ) 0.05 6* ** 0 .00 7 ) 0. 08 3* ** 0 .0 0 0 ) 0 .03 6 0 .1 52 ) 0.08 9 0 .17 5 ) 0. 06 7* ** 0 .00 2 FC 0. 00 1 0 .8 3 2 ) 0.01 3* ** 0 .00 2 ) 0. 00 7* 0 .0 9 0 0 .0 05 0. 23 6 0 .0 02 0 .7 8 8 ) 0. 00 6 0 .12 4 FC*XCash 0. 04 4* 0. 09 9 0 .0 78 ** 0.0 1 4 0 .0 9 5 * * * 0 .0 00 ) 0.0 0 9 0 .7 1 1 0 .0 9 6 0 .1 59 0. 11 3* ** 0 .00 0 Bloc k ) 0. 01 0* ** 0. 00 6 ) 0.01 0* ** 0 .00 6 ) 0. 00 9* ** 0 .0 0 7 ) 0 .01 1** * 0 .0 04 ) 0.01 2* ** 0.00 8 ) 0. 01 1* ** 0 .00 2 Block*XCash 0 .0 6 8 * * 0 .0 11 0.07 2* ** 0 .00 5 0 .0 65 ** 0.01 2 0 .08 0** * 0 .0 02 0.07 2* 0.05 2 0 .0 81 ** * 0 .0 02 No. o f O bs. 2 5 9 5 2 59 5 2 5 9 4 2 520 1 7 9 8 25 94 Adj . R 2 0. 28 3 0 .2 85 0. 28 6 0 .2 85 0 .2 9 2 0 .2 8 5 Panel B : R &D XCash 0. 02 5* ** 0. 00 2 0 .0 37 ** 0.0 0 0 0 .0 3 5 * * * 0 .0 00 0.0 0 5 0 .5 8 7 0 .0 5 7 * * 0 .0 12 0. 03 8* ** 0 .00 0 FC 0 .0 0 2 0 .3 78 0.00 1 0 .47 9 ) 0. 00 2 0 .2 34 ) 0 .00 1 0 .4 53 ) 0 .0 0 2 0 .3 18 0. 00 1 0 .58 9 FC*XCash 0. 02 5* ** 0. 00 9 ) 0.01 7 0 .15 8 ) 0. 00 1 0 .8 98 0.0 5 7 * * * 0 .0 0 0 ) 0.00 3 0 .90 3 ) 0. 01 8* 0 .09 9 Bloc k 0. 00 2* 0. 09 4 0 .0 02 0.2 1 9 0 .0 0 2 0 .1 7 7 0 .0 02 0. 16 8 0 .0 03 ** 0 .0 4 5 0 .0 0 2 0 .19 3 Block*XCash 0. 00 1 0 .9 5 2 0 .0 0 7 0 .4 19 0. 00 6 0 .4 88 0.0 0 3 0 .7 6 0 ) 0 .0 0 3 0 .8 05 0. 00 6 0 .50 8 No. o f O bs. 2 5 9 5 2 59 5 2 5 9 4 2 520 1 7 9 8 25 94 Adj . R 2 0. 37 4 0 .3 72 0. 37 2 0 .3 88 0 .3 7 4 0 .3 7 2 This table rep orts the co efficients o n the fixe d effe cts of investme nt on exces s cash unde r both fi nanc ial constraints a nd institutional b loc kholdings. T he depe ndent variable in P ane l A is capit a l expenditure (Capex ), while the depe ndent variable in Panel B is R&D expenditure (R&D ); all o f th e de pende nt vari-ables a re normaliz ed b y the start-of-year book assets. XC as h ref ers to the b egi n n ing-of-yea r exc ess cash holdings o f a firm normalized b y the start-of-ye a r book assets. FC is a dummy variable fo r fi n a ncial const raints w hic h is equal to 1 (0) if th e firm is fi nanc ially constrain ed (unc onstrained ); Bloc k is a dummy variab le w h ich is equ al to 1 (0 ) if th e institutional b lockh o ld in gs o f the fi rm are b el o w (above) the m ed ia n for the sample. The control variables includ e Cash Fl ow is cash flo w normaliz ed b y the start-of-ye a r book asse ts ; Q is the b egi n n ing-of-year m arket-to-book rati o; Size is natural lo g of assets; SalesG is th e growth in sales o ve r the p revious three-year period, and Le verag e is the sum of lo ng-term debt and curre nt liab ilities normalized by the start -o f-year book as se ts. T he p -value s b ased on robu st stan dard er rors are re ported in p are n theses. * **, * * a nd * indic a te stat is tic a l significa n ce at the 1 , 5 a n d 1 0% leve ls , re spe ctively. A ll est imation s include indu stry and y ea r ind ic ators a s w el l a s in te rc ept term. D et aile d re sults are a vailable from the authors upo n re quests.
Our paper complements recent literature by considering how both costly exter-nal finance and management entrenchment influence investment-excess cash sen-sitivity. We find that excess cash is significantly correlated with capital expenditure, particularly for firms that are financially constrained and under severe managerial entrenchment. However, excess cash is not correlated with R&D under these two dimensions.
In summary, our results have implications for corporate liquidity management in an emerging market like Taiwan. Although excess cash is beneficial for finan-cially constrained firms, it could expropriate shareholders’ interests by facilitating empire-building overinvestment when such firms also have severe managerial entrenchment. Therefore, it might be questionable for poorly managed firms to accumulate cash even though they are facing costly external funds. Further research could be carried out to find under which state excess cash can be used for value-increasing or value-decreasing investment when both financial con-straints and agency problems are considered.
References
Almeida, H., M. Campello, and M. S. Weisbach, 2004, The cash flow sensitivity of cash, Journal of Finance59, 1777–1804.
Alti, A., 2003, How sensitive is investment to cash flow when financing is frictionless?, Journal of Finance58, 707–722.
Babenko, I., M. Lemmon, and Y. Tserlukevich, 2011, Employee stock options and invest-ment, Journal of Finance 66, 981–1009.
Berger, P. G., E. Ofek, and D. L. Yermack, 1997, Managerial entrenchment and capital structure decisions, Journal of Finance 52, 1411–1438.
Brown, J. R., and B. C. Petersen, 2011, Cash holdings and R&D smoothing, Journal of Corporate Finance17, 694–709.
Callahan, W. T., J. A. Millar, and C. Schulman, 2003, An analysis of the effect of man-agement participation in director selection on the long-term performance of the firm, Journal of Corporate Finance9, 169–181.
Claessens, S., S. Djankov, and L. Lang, 2000, The separation of ownership and control in East Asia corporations, Journal of Financial Economics 58, 81–112.
Denis, D. J., and V. Sibilkov, 2010, Financial constraints, investment and the value of cash holdings, Review of Financial Studies 23, 247–269.
Dittmar, A., and J. Mahrt-Smith, 2007, Corporate governance and the value of cash holdings, Journal of Financial Economics 83, 599–634.
Dittmar, A., J. Mahrt-Smith, and H. Servaes, 2003, International corporate governance and corporate cash holdings, Journal of Financial and Quantitative Analysis 38, 111–133.
Donaldson, L., 1990, The ethereal hand: organizational economics and management the-ory, Academy of Management Review 15, 369–381.
Fama, E. F., and M. C. Jensen, 1983, Separation of ownership and control, Journal of Law and Economics26, 301–325.
Faulkender, M., and R. Wang, 2006, Corporate financial policy and the value of cash, Journal of Finance61, 1957–1990.
Fazzari, S., G. Hubbard, and B. Petersen, 1988, Financing constraints and corporate investment, Brooking Papers on Economic Activity 1, 141–195.
Florackis, C., and A. Ozkan, 2009, Managerial incentives and corporate leverage: evi-dence from the United Kingdom, Accounting and Finance 49, 531–553.
Franzoni, F., 2009, Underinvestment vs. overinvestment: evidence from price reaction to pension contributions, Journal of Financial Economics 92, 491–518.
Fresard, L., and C. Salva, 2010, The value of excess cash and corporate governance: evi-dence from US cross-listings, Journal of Financial Economics 98, 359–384.
Hadlock, C. J., and J. R. Pierce, 2010, New evidence on measuring financial constraints: moving beyond the KZ Index, Review of Financial Studies 23, 1909–1940.
Harford, J., S. Mansi, and W. Maxwell, 2008, Corporate governance and firm cash hold-ings in the US, Journal of Financial Economics 87, 535–555.
Himmelberg, C. P., and B. C. Petersen, 1994, R&D and internal finance: a panel study of small firms in high-tech industries, Review of Economics and Statistics 76, 38–51. Jensen, M., 1986, Agency costs of the free cash flow, corporate finance and takeovers,
American Economic Review76, 323–329.
Jensen, M., 1993, Presidential address: the modern industrial revolution, exit and the fail-ure of internal control systems, Journal of Finance 48, 831–880.
Kalcheva, I., and K. Lins, 2007, International evidence on cash holdings and expected managerial agency problems, Review of Financial Studies 20, 1087–1112.
Kaplan, S., and L. Zingales, 1997, Do investment-cash flow sensitivities provide useful measures of financing constraints? Quarterly Journal of Economics 112, 169–215. Khurana, I. K., X. Martin, and R. Pereira, 2006, Financial development and the cash
flow sensitivity of cash, Journal of Financial and Quantitative Analysis 41, 787–807. Kothari, S. P., T. E. Laguerre, and A. J. Leone, 2002, Capitalization versus expensing:
evidence on the uncertainty of future earnings from capital expenditure versus R&D outlays, Review of Accounting Studies 7, 355–382.
La Porta, R., F. Lopez-de-Silanes, and A. Shleifer, 1998, Law and finance, Journal of Political Economy106, 1113–1155.
La Porta, R., F. Lopez-de-Silanes, and A. Shleifer, 1999, Corporate ownership around the world, Journal of Finance 54, 471–517.
Leland, H., and D. Pyle, 1977, Information asymmetries, financial structure and financial intermediation, Journal of Finance 32, 371–387.
McLean, R. D., T. Zhang, and M. Zhao, 2012, Why does the law matter? Investor protection and its effects on investment, finance and growth, Journal of Finance 67, 313–350.
Myers, S. C., and N. S. Majluf, 1984, Corporate financing and investment decisions when firms have information that investors do not have, Journal of Financial Economics 13, 187–221.
Myers, S. C., and R. G. Rajan, 1998, The paradox of liquidity, Quarterly Journal of Eco-nomics113, 733–771.
Opler, T., L. Pinkowitz, R. M. Stulz, and R. Williamson, 1999, The determinants and implications of corporate cash holdings, Journal of Financial Economics 52, 3–46. Pinkowitz, L., and R. Willionson, 2004, What is a Dollar Worth? The Market Value
of Cash Holdings, Working paper (Georgetown University, McDonough School of Business, Washington, DC).
Pinkowitz, L., R. M. Stulz, and R. Williamson, 2006, Does the contribution of corporate cash holdings and dividends to firm value depend on governance? A cross-country anal-ysis. Journal of Finance 61, 2725–2752.
Shen, C. H., and C. A. Wang, 2005, Does bank relationship matter for a firm’s invest-ment and financial constraints? The case of Taiwan, Pacific-Basin Finance Journal 13, 163–184.
Shleifer, A., and R. W. Vishny, 1986, Large shareholders and corporate control, Journal of Political Economy94, 461–488.
Shleifer, A., and R. W. Vishny, 1989, Management entrenchment: the case of manager-specific investments, Journal of Financial Economics 25, 123–139.
Stein, J., 2003, Agency, information and corporate investment, in: G. Constantinides, M. Harris, R. Stulz, eds., Handbook of the Economics of Finance, (Elsevier/North-Holland, New York), 111–165.
Xu, N., X. Xu, and Q. Yuan, 2012, Political connection, financing frictions and corporate investment: evidence from Chinese listed family firms, European Financial Management, forthcoming.
Yeh, Y. H., and T. Woidtke, 2005, Commitment or entrenchment? Controlling sharehold-ers and board composition, Journal of Banking and Finance 29, 1857–1885.
