Agency Costs and Corporate Governance Mechanisms

Agency cost and unified judicature weapons Evidence for UK squ ars Chrisostomos Florackis and Aydin Ozkan* University of York, UK pussyfoot In this paper, we aim to ext send away the a posteriori literature on the determinants of procedure cost by using a super savour of UK listed substantials. To do so, we employ cardinal alternate proxies for function be the proportion of native sales to meat summations ( summation turn over) and the dimension of selling, general and administrative outlays (SG&A) to hit sales. In our epitome, we control for the influence of several(prenominal) infixed constitution mechanisms or devices that were do by by former studies.Also, we examine the momentiveness interactions betwixt these mechanisms and immobile exploitation opportunities in determining part be. Our ensues reveal that the outstanding organise characteristics of riotouss, viz. depository financial institution debt and debt adulthood, c all in all di ckens of the most of the essence(predicate) corporeal giving medication devices for UK companies. Also, managerial self-possession, managerial requital and possession conpennyproportionn seem to quicken an authoritative occasion in mitigating billet be. Finally, our terminuss aim that the par dispense exerted by internal memorial conf substance abuset mechanisms on say-so be varies with securelys harvesting opportunities.JEL compartmentalisation G3 G32 Keywords Agency be Growth opportunities Internal embodied Governance Mechanisms. * check author. Department of Economics and Related Studies, University of York, Heslington, York, YO10 5DD, UK. Tel. + 44 (1904) 434672. Fax + 44 (1904) 433759. wagemail email comforted ac. uk. We thank seminar participants at University of York, and the 2004 European Finance Association Meetings for serveful comments and suggestions. 1 1. Introduction Fol upseting Jensen and Meckling (1976), commission relations within the s tarchy and cost associated with them thrust been extensively finishvasd in the somatic finance literature.There is a commodious deal of trial-and-error take shape providing indorse that financial finalitys, coronation determinations and, hence, firm none prize atomic number 18 prodigiously refered by the presence of deputation conflicts and the extent of delegacy be. The instruction of these studies has been the jar of the expected post cost on the exploit of firms. 1 Moreover, the implicit assumption is that, in continuous tense expectant grocerys, way of life be arising from conflicts in the midst of firms claimholders exist and the value of firms reductions if the market expects that these cost atomic number 18 likely to be realised.It is in whatsoever face assumed that at that place argon internal and foreign merged regime mechanisms that mess armed service reduce the expected cost and their nixly charged equal on firm value. For example, practically of prior work on the willpower and performance birth relies on the s rout out that managerial possession buns queue the resides of managers and theatrical roleholders and hence wholeness would observe a confident(p) feign exerted by managerial sh atomic number 18holdings on the performance of firms. The coercive impact is argued to be referable to the decrease in the expected costs of the manner conflict amongst managers and sh atomic number 18holders.Despite much valuable insights generated by this strand of literature, however, but precise few studies at once tackle the mea incon tallyablement bit of the principal unsettled of interest, namely business office costs. Notable exceptions argon Ang et al. (2000) and cut and Davidson (2003), which investigate the observational determinants of agency costs and focus on the persona of debt and self-will grammatical construction in mitigating agency problems for the US firms. In doing so, they call ii alternate proxies for agency costs the balance of innate sales to correspond pluss (addition perturbation) and the ratio of selling, general and administrative put downs (SG&A) to total sales.In line with the materializeings of prior research they provide try out for the encounter that managerial self-possession aligns the interests of managers and shargonholders and, hence, reduces agency costs in general. However, thither is no consensus on the usance of debt in mitigating such(prenominal) problems and associated costs. Ang et al. (2000) point pop out that debt has an alleviating role whereas Sign and Davidson (2003) an aggravating one. The objective of this paper is to extend the investigation of these studies by analysing experimentally the determinants of agency costs in the UK for a broad experiment of 1See, for example, Morck et al. (1988) McConnell and Servaes (1990) and Agrawal and Knoeber (1996) among new(prenominal)s. 2 listed firms. Fol sc ummying the whole kit and caboodle of Ang et al. (2000) and, Sign and Davidson (2003), we shape both(prenominal) proxies of agency costs summation turnover and the (SG&A) ratio. More specifically, we empirically examine the impact of capital structure, possession, be on composition and managerial allowance on the costs likely to arise from agency conflicts surrounded by managers and shargonholders. In doing so, we profligately turn over particular attention to the role of ripening opportunities in influencing the authority of internal boldness mechanisms in decrease agency costs. In draw outing out the compendium in this paper, we aim to provide insights at least in trine important argonas of the empirical research on agency costs. First, in analyse the determinants of agency costs, the analysis of this paper in bodieds important firmspecific characteristics (internal somatic arrangement devices) that whitethornhap affect agency costs but were ignored by old s tudies.For example, we explore the role the debt maturity structure of firms toilet play in controlling agency costs. It is widely acknowledged that short-run debt whitethorn be to a greater extent instalive than long-term debt in reducing the expected costs of the underinvestment problem of Myers (1977). 3 Accordingly, in our analysis, we tip over the maturity structure of debt as a possible drop governing device that is effective in reducing the expected costs of the agency conflict amidst sh arholders and debtholders. Similar to Ang et al. 2000) that investigate if assert debt creates a convinced(p) externality in the form of overturn agency costs, we also check if the extraction of debt financing matters in mitigating agency problems. A nonher effectivenessly effective somatic giving medication mechanism we consider relates to managerial compensation. Repenny studies suggest that compensation contracts provide motivate managers to clear actions that maximize s h arholders wealth (see, e. g. , Core et al. , 2001 Murphy, 1999 among others). This is arsed on the opinion that financial carrots motivate managers to maximize firm value.That is, a manager will presumptively be little likely, ceteris paribus, to exert stingy effort and risk the release of his job the greater the level of his compensation. several(prenominal) empirical studies provide evidence for the effectiveness of managerial compensation as a corporate boldness mechanism. For in military posture, 2 As rationaliseed later in the paper, the devil proxies for agency costs that are open in our analysis are more than likely to stimulate the agency problems amidst managers and shareholders. However, we do non rule out the possibility that they may also mystify the agency problems amidst shareholders and debtholders. It is argued that firm with greater branch opportunities should digest more short-term debt be arrange shortening debt maturity would make it more lik ely that debt will mature before either opportunity to exercise the increment options. Consistent with this prediction, thither are several empirical debt maturity studies that run across a negative relation among maturity and return opportunities (see, e. g. , Barclay and Smith, 1995 Guedes and Opler, 1996 and Ozkan, 2000 among others). 3 Hutchinson and Gul (2004) find that managers compensation can soften the negative friendship surrounded by reaping opportunities and firm value.In this paper, we examine the effectiveness of managerial compensation as a corporate arrangement mechanism by including the sugar profit of managers in our empirical feign. We also acknowledge that there dupe been concerns more or less excessive compensation big moneys and their negative impact on corporate performance. Accordingly, we investigate the possibility of a non-monotonic impact the managerial compensation may exert on agency costs. Second, our empirical model captures potential interactions in the midst of corporate ecesis mechanisms and harvest-time opportunities.Following McConnell and Servaes (1995) and Lasfer (2002), we expect the effectiveness of governance mechanisms in reducing agency problems to be helpless on firms offshoot opportunities. In particular, if agency problems are associated with greater cultivation asymmetry (a terzetto estate problem in graduate(prenominal)-growth firms), we expect the effectiveness of corporate governance mechanisms in mitigating asymmetric t separately(prenominal)ing problems to outgrowth in high-growth firms (Smith and Watts, 1992 and Gaver and Gaver, 1993).However, if, as argued by Jensen (1986), agency problems are associated with conflicts over the use of free change current (a plebeian problem in low-growth firms), we expect governance mechanisms that are likely to mitigate such problems to play a more important role in low-growth firms (Jensen, 1986). Last but not least, in severalise to prev ious studies that focus on the US market, we provide evidence for UK firms. Although the UK and the US are usually characterized as having a standardized common fair play regulatory system (see, e. g. , La Porta et al. 1998), the UK market bears hearty distinguishing characteristics. 4 It is argued that several of these characteristics may contribute to a more foolificant tip of managerial discretion and, hence, high level of managerial agency costs. For example, notwithstanding the comparatively high proportion of shares held by financial institutions, there is a great deal of evidence that financial investors do not take an active role in corporate governance. Similarly, UK dining tables are usually characterized as corporate devices that provide weak disciplinary function.More specifically, weak fiduciary obligations on directors demand resulted in non administrator directors playing more an advisory than a monitor role. 5 Consequently, the investigation of agency issues and the effectiveness of the resource governance 4 For a more flesh out discussion about the characteristics of the prevailing UK corporate governance system see Short and Keasey (1999) Faccio and Lasfer (2000) Franks et al. (2001) and Ozkan and Ozkan (2004). 5 Empirical studies by Faccio and Lasfer (2000), Goergen and Rennebog (2001), Franks et al. 2001) and Short and Keasey (1999) provide evidence on the weak role of institutions and board of directors in reducing agency problems in the UK. 4 mechanisms in the UK, in a goal that witnesses an intensive discussion of corporate governance issues, would be of significant importance. Our results crockedly suggest that managerial monomania constitutes a strong corporate governance mechanism for the UK firms. This result is conformable with the findings provided by Ang et al. (2000) and Sign and Davidson (2003) for the US firms.Ownership slow-wittedness and pay also seem to play a significant role in mitigating agency associate problems. The results concerning the role of capital structure variables on agency costs are striking. It seems that both the source and the maturity structure of corporate debt have a significant effect on agency costs. Finally, there is strong evidence that specific governance mechanisms are not homogeneous but set out with growth opportunities. For instance, we find that executive willpower is more effective as a governance mechanism for high-growth firms.This result is complementary to the results obtained by Smith and Watts (1992), Gaver and Gaver (1993) and Lasfer (2002), which brave the view that high-growth firms are likely to prefer in pennyimeive mechanisms (e. g. managerial self-control) whereas low-growth firms focus more on monitor mechanisms (e. g. short-term debt). The re mainder of the paper is organized as follows. In section 2 we discuss the think to theory and work our empirical hypotheses. Section 3 describes the way in which we have constructed our prov e and presents several descriptive statistics of that.Section 4 presents the results of our univariate, multivariate and sensitivity analysis. Finally, section 5 concludes. 2. Agency costs and Governance Mechanisms In what follows, we will discuss the potential interactions amidst agency costs and internal corporate governance mechanisms operational to firms. Also, we will break apart how firm growth opportunities affect agency costs and the family between governance mechanism and agency costs. 2. 1 Debt back off Agency problems within a firm are usually related to free cash-flow and asymmetric study problems (see, for example, Jensen, 1986 and Myers and Majluf, 1984).It is widely acknowledged that debt servicing obligations answer reduce of agency problems of this sort. This is particularly true for the subject field of privately held debt. For example, tilling club 5 debt incorporates significant signalling characteristics that can mitigate informational asymmetry conf licts between managers and international investors (Jensen, 1986 Stulz, 1990 and Ross, 1977). In particular, the announcement of a bank credit agreement conveys despotic news to the stock market about creditors worthiness.Bank debt also bears important renegotiation characteristics. As Berlin and Mester (1992) argue, because banks are well informed and typically baseborn in number, renegotiation of a loan is easier. A banks willingness to renegotiate and re-create a loan indicates the conception of a close family family between the borrower and the creditor and that is a further good signal about the quality of the firm. Moreover, it is argued that bank debt has an advantage in comparison to publicly traded debt in supervise firms activities and in collecting and processing information.For example, Fama (1985) argues that bank lenders have a comparative advantage in minimizing information costs and getting admission charge to information not otherwise publicly available. T herefore, banks can be viewed as performing a screening role employing private information that allows them to gauge and monitor borrowers more effectively than other lenders. In addition to debt source, the maturity structure of debt may matter. For example, short-term debt may be more effective than long-term debt in reducing free cash flow problems and in signalling high quality to outsiders.For example, as Myers (1977) suggests, agency conflicts between managers and shareholders such as the underinvestment problem can be curtailed with short-term debt. Flannery (1986) argues that firms with large potential information asymmetries are likely to issue short-term debt because of the larger information costs associated with long-term debt. Also, short-term debt can be advantageous particularly for high-quality companies collect to its low refinancing risk (Diamond, 1991). Finally, if yield curve is downward sloping, issuing short-term debt increases firm value (Brick and Ravid, 1985).Consequently, bank debt and short-term debt are expected to constitute cardinal important corporate governance devices. We complicate the ratio of bank debt to total debt and the ratio of short-term debt to total debt to our empirical model so as to approximate the lenders top executive to mitigate agency problems. Also, we include the ratio of total debt to total pluss (leverage) to approximate lenders incentive to monitor. In general, as leverage increases, so does the risk of nonpayment by the firm, hence the incentive for the lender to monitor the firm6. 6 Ang et al. 2000) focus on consume of small firms, which have do not have easy access to public debt, and examine the impact of bank debt on agency costs. On the contrary, Sign and Davidson (2003) focus on a sample of large firms, which have easy access to public debt, and examine the impact of public debt on 6 2. 2 Managerial Ownership The conflicts of interest between managers and shareholders arise mainly from th e breakup between ownership and control. Corporate governance deals with finding ways to reduce the order of these conflicts and their adverse cause on firm value.For instance, Jensen and Meckling (1976) suggest that managerial ownership can align the interest between these two divergent groups of claimholders and, therefore, reduce the total agency costs within the firm. According to their model, the descent between managerial ownership and agency costs is one-dimensional and the optimal point for the firm is achieved when the managers acquires all of the shares of the firm. However, the consanguinity between managerial ownership and agency costs can be non-monotonic (see, for example, Morck et al. , 1988 McConnel and Servaes, 1990,1995 and, Short and Keasey, 1999).It has been shown that, at low levels of managerial ownership, managerial ownership aligns managers and outside shareholders interests by reducing managerial incentives for perk consumption, employ of skimpy eff ort and engagement in nonmaximizing projects (alignment effect). After some level of managerial ownership, though, managers exert inadequate effort (e. g focus on external activities), collect private benefits (e. g. build empires or enjoy perks) and entrench themselves (e. g. cut high risk projects or bend over backwards to hold a takeover) at the disbursal of other investors (entrenchment effect).Therefore the relationship between the two is non-linear. The ultimate effect of managerial ownership on agency costs depends upon the trade-off between the alignment and entrenchment effects. In the context of our analysis we propose a non-linear relationship between managerial ownership and managerial agency costs. However, theory does not shed much light on the exact nature of the relationship between the two and, hence, we do not know which of the effects will dominate the other and at what levels of managerial ownership.We, therefore, carry out a preliminary investigation about the posture of the relationship between managerial ownership and agency costs. Figure 1 presents the way in which the two variables are associated. record Figure 1 here agency costs. Our study is more similar to that of Ang et al (2000) given that UK firms use significant amounts of bank debt financing (see Corbett and Jenkinson, 1997). 7 Clearly, at low levels of managerial ownership, plus turnover and managerial ownership are stomachatoryly related. However, after managerial ownership exceeds the 10 per cent level, the relationship turns from positive to negative.A threesome turn point is that of 30 percent after which the relationship seems to turn to positive again. Consequently, there is evidence both for the alignment and the entrenchment effects in the gaffe of our sample. In assure to capture both of them in our empirical specification, we include the level, the square and the square of managerial ownership in our model as predictors of agency costs. 2. 3 Ownership Concentration A third alternative for alleviating agency problems is through strong ownership.Theoretically, shareholders could take themselves an active role in monitoring management. However, given that the monitoring benefits for shareholders are proportionate to their legality risk (see, for example, Grossman and Hart, 1988), a small or intermediate shareholder has weensy or no incentives to exert monitoring behaviour. In contrast, shareholders with substantial stakes have more incentives to supervise management and can do so more effectively (see Shleifer and Vishny, 1986 Shleifer and Vishny, 1997 and Friend and Lang, 1988).In general, the higher the amount of shares that investors hold, the stronger their incentives to monitor and, hence, comfort their investment. Although large shareholders may help in the reduction of agency problems associated with managers, they may also harm the firm by causing conflicts between large and minority shareholders. The problem usually a rises when large shareholders gain nearly to the full control of a corporation and engage themselves in self-dealing expropriation procedures at the expense of minority shareholders (Shleifer and Vishny, 1997).Also, as Gomez (2000) points out, these expropriation incentives are stronger when corporate governance of public companies insulates large shareholders from takeover threats or monitoring and the legal system does not protect minority shareholders because either of poor laws or poor enforcement of laws. Furthermore, the existence of concentrated holdings may decrease diversification, market liquidation and stocks ability to grow and, therefore, increase the incentives of large shareholders to expropriate firms resources.several(prenominal) empirical studies provide evidence consistent with that view (see, for example, Beiner et al, 2003). In order to exam the impact of ownership submerging on agency costs, we include a variable that refers to the sum of stakes of sharehold ers with right stake greater than 3 8 per cent in our regression equation. The results remain robust when the room access value changes from 3 per cent to 5 per cent or 10 per cent. 2. 4 Board of Directors Corporate governance research recognizes the essential role performed by the board of directors in monitoring management (Fama and Jensen, 1983 Weisbach, 1988 and Jensen, 1993).The effectiveness of a board as a corporate governance mechanism depends on its size of it and composition. Large boards are usually more powerful than small boards and, hence, considered necessary for organizational effectiveness. For instance, as Pearce and Zahra (1991) point out, large powerful boards help in strengthening the link between corporations and their environments, provide counsel and advice regarding strategical options for the firm and play crucial role in creating corporate identity. new(prenominal) studies, though, suggest that large boards are less effective than large boards.The ke y notion is that large boards make coordination, communication and decision-making more cumbersome than it is in smaller groups. Recent studies by Yermack, 1996 Eisenberg et al. , 1998 and Beiner et al, 2004 support such a view empirically. The composition of a board is also important. There are two components that characterize the independence of a board, the proportion of non-executive directors and the separated or not roles of chief executive officer (chief operating officer) and chairman of the board (COB).Boards with a significant proportion of non-executive directors can limit the exercise of managerial discretion by exploiting their monitoring ability and protecting their reputations as effective and autarkic decision makers. Consistent with that view, Byrd and Hickman (1992) and Rosenstein and Wyatt (1990) propose a positive relationship between the office of non-executive directors on the board and corporate performance. Lin et al. (2003) also propose a positive share pr ice reaction to the appointment of outside directors, especially when board ownership is low and the appointee possesses strong ex ante monitoring incentives.Along a slightly different dimension, Dahya et al. (2002) find that top-manager turnover increases as the calculate of outside directors increases. Other studies find exactly the opposite results. They argue that non-executive directors are usually characterized by lack of information about the firm, do not bring the requisite skills to the job and, hence, prefer to play a less confrontational role kinda than a more critical monitoring one (see, for example, Agrawal and Knoeker, 1996 Hermalin 9 nd Weisbach, 1991, and Franks et al. , 2001)7. As far as the separation between the role of CEO and COB is concerned, it is believed that separated roles can fleet to better board performance and, hence, less agency conflicts. The Cadbury (1992) report on corporate governance stretches that issue and recommends that CEO and COB shoul d be two distinct jobs. Firms should comply with the recommendation of the report for their own benefit. A decision not to combine these roles should be publicly explained.Empirical studies by Vafeas and Theodorou (1998), and Weir et al. (2002), though, which study that issue for the content of the UK market, provide results that do not support Cadburys stance that the CEO COB duality is undesirable. In the context of the UK market, UK boards are believed to be less effective than the US ones. For instance,. To test the effectiveness of the board of directors in mitigating agency problems we include three variables in our empirical model a) the ratio of the number of non-executive directors to he number of total directors, b) the total number of directors (board size) and c) a dummy variable which takes the value of 1 when the roles of CEO and COB are not separated and 0 otherwise. 2. 5 Managerial Compensation another(prenominal) important component of corporate governance is the compensation share that is provided to firm management. Recent studies by Core et al. (2001) and Murphy (1999) suggest, among others, that compensation contracts, whose use has been change magnitude dramatically during the 90s, can motivate managers to take actions that maximize shareholders wealth.In particular, as Core et al. (2001) point out, if shareholders could directly observe the firms growth opportunities and executives actions no incentives would be necessary. However, due to asymmetric information between managers and shareholders, both legality and compensation related incentives are required. For example, an increase in managerial compensation may reduce managerial agency costs in the sense that satisfied managers will be less likely, ceteris paribus, to utilize insufficient effort, perform expropriation behaviour and, hence, risk the loss of their job.Despite the central importance of the issue, only a few empirical studies examine the impact of managerial compensat ion components on corporate performance. For example, Jensen and Murthy 7 Such a result may be consistent with the governance system prevailing in the UK market given the item that UK legislation encourages non-executive directors to be inactive since it does not let down fiduciary obligations on them. Also, UK boards are dominated by executive directors, which have less monitoring power.Franks et al. (2001) confirm this view by providing evidence on a non-disciplinary role of nonexecutive directors in the UK. 10 (1990) find a statistically significant relationship between the level of pay and performance. Murphy (1995), finds that the form, rather than the level, of compensation is what motivates managers to increase firm value. In particulars, he argues that firm performance is positively related to the percentage of executive compensation that is equity based.More recently, Hutchinson and Gul (2004) analyze whether or not managers compensation can moderate the negative associat ion between growth opportunities and firm value8. The results of this study indicate that corporate governance mechanisms such as managerial remuneration, managerial ownership and non-executive directors mayhap affect the linkages between organizational environmental factors (e. g. growth opportunities) and firm performance.Finally, Chen (2003) analyzes the relationship between equity value and employees gift. He finds that the yearly stock bonus is strongly associated with the firms contemporaneous but not time to come performance. Managerial compensation, though, is considered to be a debated component of corporate governance. Despite its potentially positive impact on firm value, compensation may also work as an infectious greed which creates an environment ripe for abuse, especially at significantly high levels.For instance, remuneration packages usually include utmost(a) benefits for managers such as the use of private jet, golf club membership, enjoyment and other expenses , apartment purchase etc. Benefits of this sort usually cause severe agency conflicts between managers and shareholders. 9 Therefore, it is possible that the relationship between compensation and agency costs is non-monotonic. Similar to the case of managerial ownership, we carry out a preliminary investigation about the pattern of the relationship between salary and agency costs.As shown in figure 2, the relationship between salary and agency costs is likely to be non-linear10. In our empirical model, we include the ratio of the total salary gainful to executive directors to total additions as a determinant of agency costs. Also, in order to capture potential 8 Rather, the majority of the studies in that strand of literature turn around the causation and examine the impact of performance changes on executive or CEO compensation (see, for example, Rayton, 2003 among others). Concerns about excessive compensation packages and their negative impact on corporate performance have lead to the establishment of introductory recommendations in the form of go around practises in which firms should comply so as the problem with excessive compensation to be diminished. In the case of the UK market, for example, one of the basic recommendations of the Cadbury (1992) report was the establishment of an independent compensation committee. Also, in a posterior report, the Greenbury (1995) report, specific propositions about remuneration issues were made.For example, an issue that was stretched was the rate of increase in managerial compensation. In the case of the US market, the set of best practises includes, among others, the establishment of a compensation committee so as transparentness and disclosure to be guaranteed (same practise an in the UK) and the substitution of stock options as compensation components with other tools that promote the long-term value of the company 10 A similar preliminary analysis is carried out so as to check potential non-linearities conc erning the relationship between the rest of internal governance mechanisms and agency costs.Our results (not report) indicate that none of them is related to agency costs in a non-linear way. 11 non-linearities, we include higher ordered salary toll in the regression equation. Finally, we include a dummy variable, which takes the value of 1 when a firm pays options or bonuses to managers and 0 otherwise. Including that dummy variable in our analysis enables us to test whether or not options and bonuses themselves provide incentives to managers.As chuck (2001) points out, ignoring options is likely to incur serious problems unless managerial options are either negligible compared to ownership or almost perfectly correlated with ownership. put in Figure 2 here 2. 6 Growth Opportunities The magnitude of agency costs related to underinvestment, asset substitution and free cash flow differ significantly crossways high-growth and low-growth firms. In the underinvestment problem, mana gers may decide to pass up positive net present value projects since the benefits would mainly accrue to debt-holders.This is more severe for firms with more growth-options (Myers, 1977). Asset substitution problems, which pass along when managers opportunistically substitute higher dissension assets for low variance assets, are also more prevalent in high-growth firms due to information asymmetry between investors and borrowers (Jensen and Meckling, 1976). High-growth firms, though, face rase free cashlow problems, which occur when firms have substantial cash reserves and a tendency to set about risky and usually negative NPV investment projects (Jensen, 1986).Given the different magnitude and types of agency costs between high-growth and low-growth firms, we expect the effectiveness of corporate governance mechanisms to vary with growth opportunities. In particular, if agency problems are associated with greater underinvestment or information asymmetry (a common problem in hig h-growth firms), we expect corporate governance mechanisms that mitigate these kinds of problems to be more effective in high-growth firms (Smith and Watts, 1992 and Gaver and Gaver, 1993).However, if, as argued by Jensen (1986), agency problems are associated with conflicts over the use of free cash flow (a common problem in low-growth firms), we expect governance mechanisms that mitigate such problems to play a more important role in low-growth firms (Jensen, 1986). Several empirical studies that model company performance confirm the existence of potential interactions between internal governance mechanism and growth opportunities. For example, McConnell and Servaes (1995) find that the relationship between firm value and leverage is negative for high-growth firms and positive for low12 growth firms.Their results also indicate that equity ownership matters, and the way in which it matters depends upon investment opportunities. Specifically, they provide weak evidence that on the v iew that the allocation of equity ownership between corporate insiders and other types of investors is more important in low-growth firms. Also, Lasfer (2002) points out that high-growth firm (low-growth firms) rely more on managerial ownership (board structure) to mitigate agency problems. Finally, Chen (2003) finds that the positive relationship between yearly stock bonus and equity value is stronger for firms with greater growth opportunities.In order to capture potential interaction effects, we include interaction terms between proxies for growth opportunities and governance mechanisms in our empirical model and, also, employ sample-splitting methods (see, for example, McConnell and Servaes, 1995 and Lasfer, 2002). establish on previous empirical evidence the prediction we make is that mechanisms that are used to mitigate asymmetric information problems (free cash flow problems) are stronger in high-growth firms (low-growth firms). 3. Data and Methodology 3. 1 Data For our empir ical analysis of agency costs we use a large sample of ublicly traded UK firms over the period 1999-2003. We use two entropy sources for the compilation of our sample. Accounting data and data on the market value of equity are smooth from Datastream database. Specifically, we use Datastream to collect information for firm size, market value of equity, yearbook sales, selling general and administrative expenses, level of bank debt, short-term debt and total debt. Information on firms ownership, board and managerial compensation structure is derived from the Hemscott Guru Academic Database.This database provides financial data for the UKs top 300,000 companies, detailed data on all directors of UK listed companies, live regulatory and AFX News feeds and share price charts and trades. Specifically, we get detailed information on the level of managerial ownership, ownership concentration, size and composition of the board, managerial salary, bonus, options and other benefits. Despite the fact that data on directors are provided in a spreadsheet format, information for each item is given in a separate file. This makes data disposition for the required variables fairly complicated.For example, in order to get information about the amount of shares held by executive directors we have to combine two different files a) the 13 file that contains data on the amount of shares held by each director and b) the file that provides information about the type of each directorship (e. g. executive director vs. nonexecutive director). Also, we have to take into method of accounting the fact that several directors in the UK hold positions in more than one company. Complications also arise when we buzz off out to collect information about the composition of the board and the remuneration package that is provided to executive directors.The way in which our final sample is compiled is the undermentioned we start with a total of 1672 UK listed firms derived from Datastream. Th is number reduces to 1450 firms after excluding financial firms from the sample. After twin(a) Datastream data with the data provided by Hemscott, the number of firms further decreases to 1150. Missing firmyear observations for any variable in the model during the sample period are also dropped. Finally, we exclude outliers so as to avoid the problem with extreme values. We end up with 897 firms for our empirical analysis. 3. Dependent Variable In our analysis we use two alternative proxies to measure agency costs. Firstly, we use the ratio of annual sales to total assets (Asset Turnover) as an inverse proxy for agency costs. This ratio can be interpreted as an asset utilization ratio that shows how effectively management deploys the firms assets. For instance, a low asset turnover ratio may indicate poor investment decisions, insufficient effort, consumption of perquisites and purchase of unproductive products (e. g. office space). Firms with low asset turnover ratios are expected to experience high agency costs between managers and shareholders11.A similar proxy for agency costs is also used in the studies of Ang et al. (2000) and Sign and Davidson (2003). However, Ang et al. (2000), instead of using the ratio directly, they use the difference in the ratios of the firm with a certain ownership and management structure and the no-agency-cost base case firm. Secondly, following Sign and Davidson (2003), we use the ratio of selling, general and administrative (SG&A) expenses to sales (expense ratio). In contrast to asset turnover, expense ratio is a direct proxy of agency costs.SG&A expenses include salaries, commissions charged by agents to facilitate transactions, travel expenses for executives, advertising and marketing costs, rents and other utilities. Therefore, expense ratio should 11 The asset turnover ratio may also capture (to some extent) agency costs of debt. For instance, the sales ratio provides a good signal for the lender about how effectively t he borrower (firm) employs its assets and, therefore, affects the cost of capital 14 reflect to a significant extent managerial discretion in spending company resources.For example, as Sign and Davidson (2003) point out, management may use advertising and selling expenses to camouflage expenditures on perquisites p. 7. Firms with high expense ratios are expected to experience high agency costs between managers and shareholders12. 3. 3 Independent Variables Our empirical model includes a set of corporate governance variables related to firms ownership, board, compensation and capital structure. Several control variables are also incorporated. For example, we use the logarithm of total assets in 1999 prices as a proxy for firm size (SIZE).Also, we include the market-to- day obtain value (MKTBOOK) as a proxy for growth opportunities. Finally, we divide firms into 15 firmaments and include 14 dummy variables accordingly so as to control for sector specific effects. Analytical definitio ns for all these variables are given in circuit board 1. Insert hold over 1 here 3. 4 Methodology We examine the determinants of agency costs by employing a cross sectional regression preliminary. Following Rajan and Zingales (1995) and Ozkan and Ozkan (2004), the dependent variable is heedful at some time t, while for the independent variables we use honest-past values.Using fair(a)s in the way we construct our explanatory variables helps in mitigating potential problems that may arise due to short-term fluctuations and extreme values in our data. Also, using past values reduces the likelihood of sight relations reflecting the effects of asset turnover on firm specific factors. Specifically, the dependent variable is measured in year 2003. For accounting variables and the market-tobook ratio we use average values for the period 1999-2002. Ownership, board and compensation structure variables are measured in year 2002.Given that equity ownership characteristics in a plain ar e relatively stable over a certain period of time, we do not expect that measuring them in a sensation year would yield a significant bias in our results (see also La Porta et al. , 2002, among others). 12 An alternative proxy for agency costs between managers and shareholders, which is not used in our paper though, is the interaction of companys growth opportunities with its free cash flow (see Doukas et al. , 2002). 15 Our approach captures potential interaction effects that may be present.For example, as explained analytically in section 2. 6, the nature of the relationship between the alternative governance mechanisms or devices and agency costs may vary with firms growth opportunities. To explore that possibility, we firstly interact our proxy for growth opportunities (MKTBOOK) with the alternative corporate governance mechanisms. In this way, we test for the existence of both main effects (the impact governance variables on agency costs) and conditional effects (the impact of growth opportunities on the relationship between governance variables and agency costs).Additionally, we split the sample into high-growth and low-growth firms and estimate our empirical models for each sample separately. past we check whether the coefficients of governance variables retain their sign and their significance across the two sub-samples. 3. 5 Sample Characteristics knock back 2 presents descriptive statistics for the main variables used in our analysis. It reveals that the average values of asset turnover ratio and SG&A ratio are 1. 24 and 0. 45 respectively. The mean value for managerial ownership is 14. 4 per cent of which the average proportion of stakes held by executive (non-executive) directors is 10. 68 per cent (4. 06 per cent). The ownership concentration reaches the level of 37. 19 per cent, on average, in the UK firms. Also, the average proportion of non-executive directors is 49. 5 per cent and the average board size consists of 6. 97 directors. Finally, we were able to identify only 73 firms out of the final 897 (8. 1 per cent) in which the same person held the positions of CEO and COB. As far as the capital structure variables are concerned, the average proportion of bank debt on firms capital structure is 55. 5 per cent and that of short-term debt is 49. 53 per cent. Finally, the average market-to-book value is 2. 09. In general, these values are in line with those reported in other studies for UK firms (see, for example, Ozkan and Ozkan, 2004 and Short and Keasey, 1999). Insert board 2 here The results of the Pearsons Correlation of our variables are reported in Table 3. Our inverse proxy for agency costs, asset turnover, is clear positively correlated to managerial ownership, executive ownership, salary, bank debt and short-term debt.Ownership concentration is also positively related to asset turnover but the correlational statistics coefficient is not statistically significant. On the contrary, board size and non-executive 16 directors are prepare to be negatively correlated with asset turnover. Finally, as expected, asset turnover is found to be negatively correlated with both growth opportunities and firm size. The results for our second proxy for agency costs, SG&A, are qualitatively similar with a few exceptions (e. g. short-term debt) but with opposite signs given that SG&A is a direct and not an inverse proxy for agency costs. Insert Table 3 here 4. Empirical Results 4. 1 Univariate analysis In Table 4 we report univariate mean-comparison test results of the sample firm subgroups categorized on the basis of above and below median value values for managerial ownership, ownership concentration, board size, proportion of non-executives, bank debt, short-term debt, total debt, salary, firm size and growth opportunities. Firms with above median managerial ownership (ownership concentration) have asset turnover of 1. 34 (1. 31) whereas those with below median managerial ownership (ownership concent ration) have asset turnover of 1. 5 (1. 17). These differences are statistically significant at the 1 per cent (5 per cent) level. The results for executive ownership, salary, bank debt and short-term debt are also found to be statistically significant and are in the hypothesized direction. Specifically, we find that firms with above median values for all the above mentioned variables have relatively higher asset utilization ratios. On the contrary, there is evidence that firms with larger board sizes indicate significantly lower asset utilization ratios. Insert Table 4 here In panel B of the same table we report the results using SG&A expense ratio as a proxy for agency costs. Results are in general not in line with the hypothesized signs with notable exceptions those of ownership concentration and growth opportunities. For example, firms with above median ownership concentration (MKTBOOK) have an SG&A expense ratio of 0. 41 (0. 55) whereas firms with below median ownership concen tration (MKTBOOK) have an SG&A expense ratio of 0. 49 (0. 36).However, the results for managerial ownership, salary and short-term debt suggest that these governance mechanisms or devices are not effective in protecting firms from excessive SG&A 17 expenses. Sign and Davidson (2003) obtains a set of similar results, for the case when agency costs are approximated with the SG&A ratio. Overall, the univariate analysis indicates several corporate governance mechanisms or devices, such as managerial ownership, ownership concentration, salary, bank debt and short-term debt, which can help mitigate agency problems between managers and shareholders.Also, consistent with previous studies, we find that the relation between governance variables and agency costs is stronger for the asset turnover ratio than the SG&A expense ratio. The analysis that follows allows us to test the validity of these results in a multivariate framework. 4. 2 multivariate analysis In this section we present our resu lts that are based on a cross sectional regression approach. We start with a linear specification model, where we include only total debt from our set of capital structure variables (model 1).In general, the estimated coefficients are in line with the hypothesized signs. Specifically, consistent with the results of Ang et al. (2000) and Sign and Davidson (2003), we find both managerial ownership and ownership concentration to be positively related to asset-turnover. The coefficients are statistically significant at the 5 per cent and 1 per cent significance level respectively. On the contrary, the coefficient for board size is negative, which credibly indicates that firms with larger board size are less efficient in their asset utilization.Also, the results for our proxy for growth opportunities (MKTBOOK) support the view that high-growth firms suffer from higher agency costs than low-growth firms. Finally, there is strong evidence that managerial salary can work as an effective inc entive mechanism that helps aligning the interests of managers with those of shareholders. Specifically, the coefficient for salary is positive and statistically significant to the 1 per cent level. Therefore, compared to previous studies, our empirical model provides evidence on the existence of an additional potential corporate governance mechanism available to firms. Insert Table 5 here In model 2 we incorporate two additional capital structure variables, the ratio of bank debt to total debt and the ratio of short-term debt to total debt, in order to test whether debtsource and debt-maturity impacts agency costs. Also, we split managerial ownership into executive ownership (the amount of shares held by executive directors) and non-executive 18 ownership (the amount of shares held by non-executive directors). We do this because we expect that equity ownership works as a better incentive mechanism in the turn over of executive directors rather in the hands of non-executive direct ors.According to our results, bank debt is positively related to asset turnover. Also, in addition to debt source, the maturity structure of debt seems to have a significant effect on agency costs. The coefficient of short-term debt is positive and statistically significant at the 1 per cent significance level. Furthermore, there is evidence that from total managerial ownership, only the amount of shares held by executive directors can enhance asset utilization and, hence, align the interest of managers with those of shareholders.In model 3 we estimate a non-linear model by adding the square of salary. As explained earliest in the paper, a priori expectations, which are supported by preliminary in writing(p) investigation, suggest that the relationship between asset turnover and salary can be non-monotonic. Our results provide strong evidence that the relationship between salary and asset turnover is non-linear. In particular, at low levels of salary, the relationship between sala ry and asset turnover is positive. However, at higher levels of salary, the relationship becomes negative.This result is consistent with studies that suggest that extremely high levels of salary usually work as an infectious greed and create agency conflicts between managers and shareholders. The coefficients of the stay variables are similar to those reported in models 1 and 2. Finally, in model 4 we allow for a non-linear relationship between executive ownership and agency costs. However, our results do not support such a relationship and, therefore, the square term in our following models13.To sum up, the results of Table 5 indicate that managerial ownership (executive ownership), ownership concentration, salary (when it is at low levels), bank debt and short-term debt can help in mitigating agency problems by enhancing asset utilization. Also, the coefficients for the control variables market to book and firm size, negative and positive respectively, suggest that smaller and non - growth firms are associated with reduced asset utilization ratio and, hence, more severe agency problems between managers and shareholders.As discussed earlier in the paper, there is a possibility that the nature of the relationship between the alternative governance mechanisms or devices and agency costs varies with firms growth opportunities. In Panel A of Table 6, we explore such a In trial regressions, which are not reported, the cubic term of executive ownership is also included in our model. Once more, the results do not support the existence of a non-monotonic relationship. 13 19 possibility by interacting those governance mechanisms found significant in models 1-4 with growth opportunities, proxied by market-to-book ratio.Our empirical results support the existence of two interaction effects. We find that executive ownership is an effective governance mechanism especially for high-growth firms (the coefficient EXECOWNER* MKTBOOK is positive and statistically significant). This result is consistent with the study of Lasfer (2002), which suggests that the positive relationship between managerial ownership and firm value is stronger in high-growth firms. On the contrary, the coefficient SHORT_DEBT*MKTBOOK is found to be negative and statistically significant.This means that the qualification of short-term debt in mitigating agency problems is lower for high-growth firms. A possible explanation may be that short-term debt fundamentally mitigates agency problems related to free cash flow. Given that high-growth firms do not suffer from severe free cash-flow problems (but mainly from asymmetric information problems), the efficiency of short-term debt as governance device decreases for these firms. One could argue, though, that short-term debt should be more important for the case of highgrowth firms since it helps reduce underinvestment problems.However, it seems that this effect is not very strong for the case in our sample. A similar result is obtained in McConnell and Servaes (1995) who find that the relationship between corporate value and leverage is positive (negative) for low-growth (high-growth) firms14. Insert Table 6 here Secondly, we use the variable MKTBOOK so as two split the sample into two subsamples. We label the upper 45 per cent in terms of MKTBOOK as high-growth firms and the lower 45 per cent as low-growth firms. Then, we re-estimate our basic model for the two sub-samples separately (Table 6, panel B).The results of this exercise confirm the existence of an interaction effect between executive ownership and asset turnover. In particular, the coefficient of EXECOWNER is positive and statistically significant only in the case of the sample that includes only high-growth firms. As far as short-term debt is concerned, it is found to be positive and statistically significant in both samples. 14 The idea in McConnell and Servaes (1995) is that debt has both a positive and a negative impact on the value of the firm bec ause of its influence on corporate investment decisions.What possibly happens is that the negative effect of debt dominates the positive effect in firms with more positive net present value projects (i. e. , high-growth firms) and that the positive effect will dominate the negative effect for firms with fewer positive net present value projects (i. e. , low-growth firms). 20 To summarize, the results of our multivariate analysis suggest, among others, that executive ownership and ownership concentration can work as effective governance mechanisms for the case of the UK market.These results are in line with the ones reported by the studies Ang et al. (2000) and sign and Davidson (2003). Also, we find that, in addition to the source of debt, the maturity structure of debt can help to reduce agency conflicts between managers and shareholders. The fact that previous studies have ignored the maturity structure of debt may partly explain their contradicting results concerning the relation ship between capital structure and agency costs. Furthermore, we find that salary can work as an additional mechanism that provides incentives to managers to take valuemaximizing actions.However, its impact on asset turnover is not always positive i. e. the relationship between asset turnover and salary is non-monotonic. Finally, there is strong evidence that the relationship between several governance mechanisms and agency costs varies with growth opportunities. Specifically, our results support the view that the positive relationship between executive ownership (short-term debt) is stronger for the case of high growth (low growth) firms. 4. Robustness checks Given the significant impact of growth opportunities on agency costs (main impact) and on the impact of other corporate governance mechanisms (conditional impact), we further investigate the relationship between growth opportunities, governance mechanisms and agency costs. At first, we substitute the variable MKTBOOK with an a lternative proxy for growth opportunities. The new proxy is derived after employing common factor analysis, a statistical technique that uses the correlations between observed variables to estimate common factors and the structural relationships linking factors to observed variables.The variables which are used in order to isolate latent factors that account for the patterns of colinearity are following variables MKTBOOK = Book value of total assets minus the book value of equity plus the market value of equity to book value of assets MTBE = Market value of equity to book value of equity METBA = Market value of equity to the book value of assets METD = Market value of equity plus the book value of debt to the book value of assets. 21 These variables have been extensively used in the literature as alternative proxies for growth opportunities and Tobins Q.As shown in Table 7 (panel A) all these variables are highly correlated to each other. In order to make sure that principal compone nt analysis can provide valid results for the case of our sample, we perform two tests in our sample, the Barletts test and the Kaiser-Meyer-Olkin test. The first test examines whether or not the intercorrelation matrix comes from a population in which the variables are noncollinear (i. e. an identity matrix). The second test is a test for sampling adequacy.The results from these tests, which are reported in panel B, are encouraging and suggest that common factor analysis can be employed in our sample since all the tetrad proxies are likely to measure the same intimacy i. e. growth opportunities. Panel C presents the eigenvalues of the reduced correlation matrix of our four proxies for growth opportunities. Each factor whose eigenvalue is greater than 1 explains more variance than a single variable. Given that only one eigenvalue is greater than 1, our common factor analysis provides us with one factor that can explain firm growth opportunities.Clearly, as shown in panel D, the fa ctor is highly correlated with all MKTBOOK, MTBE, METBA and METD. We name the new variable development and use it as an alternative proxy for growth opportunities. Descriptive statistics for the variable ripening are presented in panel D. Insert Table 7 here Table 8 presents the results of cross-section analysis after using the variable GROWTH as proxy for agency costs. In general, the results of such a task are similar to the ones reported previously.For instance, there is strong evidence that executive ownership, ownership concentration, salary, short-term debt and, to some extent, bank debt are positively related to asset turnover. Also, there is some evidence supporting a non-linear relationship between salary and asset turnover. Finally, our results clearly indicate that agency costs differ significantly across high-growth and low-growth firms and, most importantly, there is a significant interaction effect between growth opportunities and executive ownership.However, we can n ot provide any evidence on the existence of an interaction between asset turnover and short-term debt. Insert Table 8 here 22 In panel B of table 8, we split our sample into high-growth and low-growth firms on the basis of high and low values for the variable GROWTH. Specifically, we label the upper 45 per cent in terms of GROWTH as high-growth firms and the lower 45 per cent as low-growth firms. Then we estimate our basic model for each sub-sample separately. The results are very similar to the ones reported in Table 6 (panel B), where we apply a similar methodology.As an additional robustness check, we use a third proxy for growth opportunities, a dummy variable that takes the value of 1 if the firm is a high-growth firm and 0 otherwise, and re-estimate the models 6 and 7 of Table 8. The definition used in order to distinguish between high-growth and low-growth firms is the following Firms above the 55th percentile in terms of the variable GROWTH are called high-growth firms. Firm s below the 45th percentile in terms of the variable GROWTH are called low-growth firms.Finally, firms between the 45th and 55th percentile are excluded from the sample. The results (not reported) are qualitatively similar to the ones reported in Table 8. For example, there is evidence for the existence of an interaction effect between executive ownership and growth opportunities but not for the one between short-term debt and growth opportunities. Also, we re-estimate the models reported in Table 8 after substituting the total salary paid to executive directors for the total remuneration package paid to executive directors.We are doing so given that the total remuneration package that is paid to managers includes several other components. For instance, the components of compensation structure have been change magnitude in number during the last decade and may include annual performance bonus, fringe benefits, stock (e. g. preference shares), stock options, stock gustatory sensati on rights, phantom shares and other deferred compensation mechanisms like qualified retirement plans (see kill and Perry, 2003 for an analytical discussion). Once more, the results do not change substantially.Finally, in Table 9 we substitute the annual sales to total assets with the ratio of SG&A expenses to total sales. As already mentioned earlier in the paper, this ratio can be used as a direct proxy for agency costs. Our results, as presented in Table 9, indicate that executive ownership, ownership concentration and total debt help reduce discretionary spending and, therefore, the agency conflicts between managers and shareholders. Sign and Davidson (2003) do not find any evidence to support these results. Also, we find that agency costs and growth opportunities are positively related i. . the coefficient of the variable GROWTH is positive and statistically significant to the 5 per cent statistical level. 23 Finally, our results support the existence of an interaction effect b etween growth opportunities and executive ownership. However, once more, our analysis does not indicate the existence of an interaction effect between short-term debt and growth opportunities. Insert Table 9 here 5. Conclusion In this paper we have examined the effectiveness of the alternative corporate governance mechanisms and devices in mitigating managerial agency problems in the UK market.In particular, we have investigated the impact of capital structure, corporate ownership structure, board structure and managerial compensation structure on the costs arising from agency conflicts mainly between managers and shareholders. The interactions among them and growth opportunities in determining the magnitude of these conflicts have also been tested. Our results strongly suggest managerial ownership, ownership concentration, executive compensation, short-term debt and, to some extent, bank debt are important governance mechanisms for the UK companies.Moreover, growth opportunities is a significant determinant of the magnitude of agency costs. Our results suggest that highgrowth firms face more serious agency problems than low-growth firms, possibly because of information asymmetries between managers, shareholders and debtholders. Finally, there is strong evidence that some governance mechanisms are not homogeneous but vary with growth oppo