Financial Study on Financial Decisions and Growth Opportunities

Description
Financial Study on Financial Decisions and Growth Opportunities: Spanish Firm's Panel Data Analysis:- Modern capital markets are almost invariably hosted on computer-based electronic trading systems; most can be accessed only by entities within the financial sector or the treasury departments of governments and corporations, but some can be accessed directly by the public. There are many thousands of such systems, most only serving only small parts of the overall capital markets.

Financial Study on Financial Decisions and Growth Opportunities: Spanish Firm's Panel Data Analysis
This paper analyses the in?uence of financial leverage decisions, dividend payout policies and the ownership structure on the firm market value when companies either face, or do not face, profitable growth opportunities. A sample of 101 large nonfinancial publicly-traded Spanish companies is used. The results confirm the relevance of debt and dividends in terms of firm value creation by showing a negative relationship between firm value and both leverage and dividend payments in the presence of growth opportunities. On the contrary, this relationship turns out to be positive when firms have no profitable investment projects. The results also demonstrate the relevance of ownership structure in the allocation of firm resources.

I. Introduction The in?uence of leverage and dividends on firm value has been a traditional topic that both academics and practitioners have paid much attention to. In a world of frictionless markets, leverage and dividends are irrelevant, in terms of firm market value, as long as they do not alter the set of firm investment opportunities at the firm disposal (Modigliani and Miller, 1958; Miller and Modigliani, 1961). On the contrary, in an imperfect market framework, the irrelevance propositions no longer held. The evolution of corporate finance in the last 40 years can be understood as the process to introduce market imperfections - basically transaction costs and taxes - in this analysis benchmark. Recently, new perceptions about the nature of debt (types and maturity) along with its impact and

that of dividends on the problems arising from the stakeholders' interest con?ict have provided new answers (Jensen, 1986; Barclay and Smith, 1996). Both decisions afect the agency relationships in two ways: (i) according to the agency explanation, lever- age and dividends modify the interest con?ict among the cash ?ow claimholders and (ii) according to the asymmetric information explanation, both decisions convey information to capital markets, mitigating adverse selection problems (Harris and Raviv, 1991; Miller and Rock, 1985). Underlying this approach is a deep redefinition of corporate financial decisions, so that the interrelation of all these topics becomes more and more important. The independence among financial decisions is no longer accepted and the optimal value-maximizing combination is intended to be found. However, given the manager discretionality and the control

* Corresponding author. E-mail: [email protected]
Applied Financial Economics ISSN 0960-3107 print/ISSN 1466-4305 online # 2005 Taylor & Francis Group Ltd http://www.tandf.co.uk/journals DOI: 10.1080/09603100500039201 391

392
systems limitations, it is not unusual that managers seek to maximize their own utility, even at the expense of shareholder's wealth. There is, in turn, a relationship between fund sources and investment, that holds both when firms face positive NPV opportunities and when they do not. In the presence, or absence, of growth opportu- nities, the ownership and control structure also play an important role in reducing the above mentioned agency problems. Although corporate governance mechanisms can reduce the interest con?ict in both situations, its role could be more important when there are no growth opportunities since undertaking unprofitable projects or perquisite consumption might exacerbate agency problems. Firm value, finan- cial structure, corporate governance, growth oppor- tunities and dividends policy then become more closely related. The aim of this study is to disentangle, at least partially, those relationships in face of the presence, or absence, of growth opportunities from a sample of Spanish companies during the 1991-1995 period. This research follows that of Andre´s et al. (2000) and draws also on the contributions of Myers (1977), Jensen (1986), Morck et al. (1988), Stulz (1990), Smith and Watts (1992), Lasfer (1995) and, very heavily, on McConnell and Servaes (1995). These last two authors are among those proposing to sort out companies according to their growth opportunities using variables like price earning ratio (PER), or the market-to-book ratio (Smith and Watts, 1992; Lasfer, 1995; McConnell and Servaes, 1995). However, the present study deviates from that research by focusing, not only on debt in?uence, but also on another strategic financial decision (dividend policy) in order to expand the analysis framework. Dividend policy has been considered a disciplinary mechanism as long as it allows for the releasing of resources when a firm has no profitable projects and, at the same time, conveys information about a firm's future expectations to capital markets. Therefore, in addition to specifying the efects on firm market value, the simultaneous consideration of leverage and dividends permits us to know their interrelationships more in depth and to shed some light on the possible complementarity or substitutability of both decisions. The results show that leverage, dividends and ownership structure remarkably afect firms' value; the kind of in?uence depending on the presence or absence of investment opportunities. When firms have positive growth opportunities debt has a negative in?uence on market value, whereas when firms do not have growth opportunities, control mechanism are more necessary so that debt and

P. de A. Alonso et al.
dividends become complementary - but not excluding - mechanisms to deal properly with manager discretionality. The ownership structure also comes out related to firm market value. To achieve these goals the paper is divided into five sections, this introduction being the first one. Section II surveys previous research, presents theoreticalfoundations of the work and introduces the hypoth- esis that the study will try to test. In Section III, some methodological issues can be found, along with the sample and variables description, while Section IV displays and comments on the results achieved and reports a sensitivity analysis to alternative specifica- tions of the model. The final section draws some conclusions from the most outstanding results and points out some future research directions that the paper proposes.

II. Theoretical Foundations Debt, dividends and growth opportunities As stated above, the existence or lack of profitable growth opportunities afordable by the com- pany in?uences the managers-shareholders interest con?ict. In order to shed some light on this contro- versy one considers the role that debt, dividends and corporate governance structure play both in the presence, and in the absence, of profitable projects. In the first scenario, the underinvestment problem is likely to arise (Myers, 1977). In essence, and as it is widely known, the underinvestment problem stresses the shortcoming of excessive debt financing in the presence of growth opportunities since too much debt can prevent managers from undertaking positive NPV projects. If this is the case, under the pressure of high financial leverage ratios, managers, acting on behalf of the shareholders, may forgo some profitable projects. The rationality underlying this fact is the priority bondholders have over firm cash ?ows relative to shareholders. If debtholders are the prior claimholders, managers do not find it worthwhile undertaking investment projects whose cash ?ows will not be perceived by company owners but by creditors. The consequences over firm value of this behaviour are clear, so that a decrease in the value can be expected due to the missing of profitable opportunities. In order to mitigate this problem, growth opportu- nities should be financed with equity instead of debt. As Myers (1977) and McConnell and Servaes (1995) assert, the higher the growth opportunities are set, the lower the leverage rate should be or, in other terms, a negative relationship between debt and firm value in

Financial decisions and growth opportunities
the face of profitable opportunities is forecast. The capitalization process consists of not only using equity rather than debt, but also of cutting dividend payments to raise the resources needed to fund invest- ment opportunities (Fazzari et al., 1988; Lang and Litzenberger, 1989; Gonza´lez, 1995). Hence, the chances are that a negative relationship between firm value and dividend payments holds if a company has profitable projects.1 The second scenario, defined by the absence of growth opportunities and closely related to the free cash ?ow hypothesis and the overinvestment prob- lem, has been the core of a number of recent researches (Jensen, 1986, 1993; Smith and Watts, 1992; Lasfer, 1995; McConnell and Servaes, 1995; Lang et al., 1996). The free cash ?ow hypothesis underlines the negative consequences of an excessive amount of resources within the reach of the managers after financing the positive NPV investment projects, especially when firms no longer have profitable growth opportunities. The free cash ?ow should be reduced by issuing new debt or paying dividends. Otherwise, it would be wasted in inefcient uses. Consequently, when firms have too much free cash ?ow and no investment opportunities, dividends and debt will probably have a positive efect on firm value. In order to test, empirically, this idea, a second hypothesis is formulated according to when compa- nies lack growth opportunities, manager disciplinary mechanisms - like debt or dividends - must be posi- tively correlated with firm value. As in the previous case, the appropriate interpre- tation of the mixed in?uence of debt and dividends on firm value requires to take into account some caveats. This is due not only to the fact that both decisions may be substitutable - either of them being able to become useless if the other one is fully used - but also due to the interaction between both of them. The dividend policy may provide evidence of the shareholders versus bondholders con?ict as long as a high, or low, enough payout ratios could give rise to wealth transfers to the first or second group of claimholders, respectively (Smith and Warner, 1979). From this point of view, debt and dividend decisions may enhance firm value creation by

393
reducing manager discretionary behaviour at the same time that are the consequence of a trade-of between shareholders' and creditors' rights. To some extent this could distort the interpretation of the results.

Ownership structure and growth opportunities Debt and dividend decisions are not the only determinants of firm value because ownership structure can have a significant in?uence too. The separation between ownership and control, so widely spread in most of the companies, causes some agency problems that, unless being properly dealt with by the external corporate control mechanisms, demand a more active role of the owners of the firm (Fama and Jensen, 1983). This can be inferred from Demsetz (1983), Demsetz and Lehn (1985), Stiglitz (1985), Jensen (1986), Shleifer and Vishny (1986) Bergstro¨m and Rydqvist (1990), and McConnell and Servaes (1990), who summarize previous research analysing the relationship between ownership structure and firm results on top of the in?uence of ownership structure on the resolution of agency problems between owners and managers. One of the most outstanding issues in this set of relationships is the proportion of the ownership in the hands of the managers because a higher proportion can make the interest of shareholders and managers to converge (Jensen and Meckling, 1976; Leland and Pyle, 1977). So, the higher the participation of man- agers in the firm ownership, the more efcient their behaviour, and a positive relationship between firm value and managers ownership is likely to hold.2 In addition, managers' participation in firm ownership can be understood in capital markets as a signal conveyed in order to show the managers' reliance on the firm investment projects (Morck et al., 1988; McConnell and Servaes, 1990). However, managers' participation is not the only way to align shareholders' and managers' interests. Stiglitz (1985) and Jensen (1986) suggest to concen- trate the ownership in the hands of a few share- holders since these have more incentives to monitor managers' work. Otherwise, in a widely dispersed

1

Notwithstanding, both debt and dividends can be used as signalling mechanisms to convey good firm investment expectations to capital markets (Bhattacharya, 1979; Campbell, 1979). Firms with the most profitable opportunities, relying on their future cash ?ows, could display higher leverage or dividend payout ratios in order to persuade the investors about their good prospects. 2 This inference may appear as too simple since some authors have shown how a linear and positive relationship between firm value and managers' ownership proportion does not hold (Morck et al., 1988; Stulz, 1988). A non-linear relationship seems to be more plausible, combining the alignment (positive relationship) and entrenchment (negative) hypothesis. This issue is dealt with in the sensitivity analysis section.

394
ownership structure, the free-rider problem arises due to the unbalanced trade-of between the efort required and the benefits the monitoring task entails. Notwithstanding, an excessive concentrated owner- ship can produce adverse consequences since it can become an obstacle when the firm faces profitable growth opportunities demanding the ownership and control specialization (Burkart et al., 1997). Hence, ownership concentration may originate two possible efects: on the one hand, it reduces agency problems by enhancing a more in-depth control and, on the other hand, it could prevent growth opportunities' exploitation. This section ends with a brief comment about the main shareholder nature. The underlying intuition is that financial intermediaries seem to be more suitable to monitor and control manager discretionality given their emphasis in producing and channelling reliable information about borrowers (Bhattacharya and Thakor, 1993). As a result, it is quite probable that firm value will be higher when the main shareholder is a bank or another financial intermediary, 3 but a priori diferences depending on the growth oppor- tunities cannot be defined. After having surveyed the theoretical foundations, let us focus on a sample of large Spanish companies in order to test the impact of financial decisions and ownership structure on firm value conditional on growth opportunities' availability. Six hypotheses are proposed: (1 and 2) Debt and dividends nega- tively afect firm value when companies have growth opportunities. (3 and 4) Debt and dividends have a positive in?uence on firm value when there are not profitable growth opportunities. (5) The dual role of investors being, at the same time, directors or man- agers is a significant determinant of value creation. And (6) ownership concentration should have a significant efect on firm value except when investment projects require a specialized ownership and control structure. In addition, the test of hypotheses 1 to 4 can be useful to analyse the substitutable versus complementary role of debt and dividends in order to solve underinvestment and overinvestment problems. III. Research Design Sample

P. de A. Alonso et al.

The sample includes 101 non-financial Spanish companies publicly traded in capital markets for the 19911995 period. Combining the 101 companies for five years we have formed a 505-observations balanced panel data which will be dealt with by the appropriate panel data methodology. Although the number of companies is not too high - the sample accounts for just about half of the Spanish quoted companies4 - the included companies are the most important ones. The sample accounts for between 72% and 80% of quoted companies capitalization and assets value is, on average, 66.13% of all quoted companies' assets. The sample selection process has been led by mar- ket data significance. In the Spanish stock market there are a very large number of quoted companies whose shares are not traded but a few days every year. It means that, in spite of being quoted compa- nies, the price of the shares do not fully re?ect future expectations. Therefore, the selected 101 companies are those more often quoted.5 From the study's point of view, the number of the companies in the sample should not be considered as a shortcoming of the study since the analysed companies are the most representative of Spanish capital markets ones. In any case, the importance of market data would be stressed because growth opportunities' identification is critically afected by the market as a benchmark and this is why investors' judgement must be expli- citly taken into account. The source of information has been the Comisio´n Nacional del Mercado de Valores (Spanish Stock Exchange Commission), hereinafter CNMV. All the data were publicly available and were obtained from the Companies Register, the Significant Ownership Participation Register and from the Audited Financial Statements.6 The CNMV provides financial information about non-financial companies and financial statements had to be complemented with ownership structure data (proportion of shares owned by directors, ownership concentration and

3

This assertion not only concerns a financial intermediary but it also may be extended to other non-financial companies like multinational firms or non-financial domestic firms. In any case, we would like to underline the more specialized control this kind of main shareholders could exercise. 4 This proportion ranges from 44% to 54% depending on the year. 5 In Spain there are two kinds of stock markets depending of the trade frequency: the mercado continuo (continuous market) and the mercado de corros (ring market). Basically, the sample comprises the companies in the mercado continuo along with the most often traded companies in the mercado de corros. 6 The original name of the databases are the Registro de Empresas, the Registro de Participaciones Significativas en el Capital and the Estados Financieros Auditados.

Financial decisions and growth opportunities
information about the main shareholder status). Given that ownership data disclosure in Spain is more constraining and it is reported only in the Significant Ownership Participation Register, a num- ber of companies had to be dropped, so the sample was reduced up to a final balanced panel data with 101 companies. Certainly, the companies in the sample are basi- cally medium to large companies compared with the average Spanish firm size either in terms of assets, sales or employees. This could raise some caveat about a possible sample bias. Notwithstanding, as Table 1's descriptive statistics show, firm size (in terms of assets) is quite heterogeneous and highly dispersed around the mean value, so the results are not supposed to be biased by size issues. The sample composition is quite industrybalanced, although there is a slight bias towards Building firms at the expense of Trade and retailing companies that canbe explained by the heavier concentration overweight of the former in the Spanish market. One should take into account that the Spanish corporate system has much in common with European corporate governance models and does not show so much ownership and control specialization as the Anglo-Saxon one. In Spanish companies, like in other European countries, ownership is more concen- trated (Berglof, 1990; Allen and Gale, 1994; Andre´s and Lo´pez, 1997) there are significant blockholders (Becht and Ro¨ell, 1999) and banks play an active role in funding and monitoring (Prowse, 1994).

395
More specifically, Spanish corporate systems could be defined by three features: (1) A high percent- age of shares owned by main shareholders, which implies a majority control such as that of France, Germany or Italy and diferent from the US system (Berglo¨f, 1990; Prowse, 1994; La Porta et al., 1999). (2) The importance of blockholders (23.76% of the companies have a multinational firm and 23.17% have a family as the main shareholder). (3) The outstanding fraction of shares owned by corporate board directors. These characteristics mean a lower ownership and control separation compared to Anglo-Saxon companies. On the one hand, agency problems stem- ming from ownership and control separation could be smaller than US companies. But, on the other hand, some problems such as risk concentration, the forgoing of specialization advantages (managers ability, specific investment, etc.) in face of profitable growth opportunities (Burkart et al., 1997) or minority shareholders expropriation (La Porta et al., 1998) could arise.

Variables The available data were intended to comprise a number of features of the companies as the existence or absence of valuable growth opportunities, capital structure, dividend payout policy, ownership and control structure and market valuation. In the appen- dix a list of all the variables and how they have been

Table 1. Descriptive statistics Descriptive statistics for the 1991-1995 and 101 Spanish firms panel data. Data about the main shareholder nature are also reported in Table 2 along with ownership variables. Assets in millions of pesetas. 1E 166866 pesetas (m) stands for market values whereas (b) stands for book values. Mean DTA (b) DTA (m) DBDT 0.4576 0.4960 0.4068 0C1 (%) 98.5 54.99 23.9857 20.6589 10.2014 0.2198 0.2376 0.2317 0.1941 0.1406 0LOGMV 6.587 0.941 0.0122 1.0556 139 939.5 Std. dev. 0.2529 0.2565 0.2746 44.3469 0.011 C2 (%) 99.27 69.89 25.3130 18.0505 0.4145 0.4260 0.4223 0.3959 0.3479 4.5206 2.080 Q 3.643 0.2258 0.8258 421 444.3 Median 0.4617 0.511 0.419 25.8327 53.8924 0.011 C5 (%) 99.69 8.03 0.89 0 0 0 0 0 0.7258 1.0360 0.213 0.0029 0.854 28 063 Max. 0.965 0.960 0.964 42.28 25.0461 61.3461 0.011 96.3 89.5 1 1 1 1 1 4.480 0.4556 0.189 8.096 3 909 311 Min. 0.005 0.010

ALFA (%) ADJALFA (%) DOM MULT FAM BA STAT

0 0 0 0 0 0

DIVTA MB Assets

0 0 558

396
constructed can be found, whereas Table 1 displays some of their basic statistics. Now let us describe brie?y the most important issues related to the specification of the variables. A key aspect of the study is the identification of the availability of growth opportunities, so that the choice of the way to measure that feature becomes crucial. The PER (price-earning ratio)7 has been chosen. There is a general agreement that this variable is a good indicator of future growth oppor- tunities by incorporating the market point of view about the firm ability to generate cash ?ows in the future (Smith and Watts, 1992; Lang and Stuz, 1994; Berger and Ofek, 1995). PER is positively related to growth opportunities, so that the higher the PER, the lower the equity value due to assets-in-place and, in turn, the higher the impact of growth opportunities on firm value (Chung and Charoenwong, 1991). As a consequence of this reason, the sample was split into two sub-samples (firms with or without profit- able growth opportunities) according to McConnell and Servaes' (1995) procedure by dividing the whole number of firms into three groups as a function of the PER value. Those companies in the upper third are certain to have more growth opportunities, while those in the lowest third could be quite reasonably characterized by the lack of valuable projects. As far as capital structure is concerned, the debt- tototal asset ratio (DTA)8 has been chosen while dividend policy has been measured by the dividend payments over total assets ratio (DIVTA). Firm mar- ket valuation was proxied by an indicator of value

P. de A. Alonso et al.
creation as the financial q (Q) or the asset marketto-book ratio (see the variables glossary in the appen- dix for a more systematic definition of all the variables and Table 1 for some descriptive statistics). Regarding the governance structure, the ALFA variable has been defined as the proportion of shares owned by the members of the board of directors. This variable was later redefined in order to take into account the directorships held by ordinary people and to exclude banks, firms and other legal entities, obtaining a more accurate proxy of the incentives directors have to run the company more efciently: the so-called adjusted-ff (ADJALFA). The ownership concentration was measured by the proportion of the total number of shares held by the main (C1), the two main (C2) and the five main (C5) share- holders. These variables can show a majority control (C1 equals 44.346% as displayed in Table 2) and proxy the extent of ownership and control specializa- tion. A brief overview of the equity concentration and the ownership distribution among shareholders types can be found in Table 2. Nevertheless, C1 may not be an informative enough indicator, so a set of five complementary dummy variables was defined to describe the nature of the main shareholder: STAT for State, MULT for a multinational firm, BA for a bank, DOM for other domestic firm and FAM for a family or a private individual or group. This classification may make sense since managers monitoring and control relies heavily on the expertise, experience and incentives of the main shareholder. 9

Table 2. Corporate ownership descriptive statistics Equity concentration and ownership distribution among diferent shareholders status. Average C1 ALFA ADJALFA % of companies Domestic 45.71 20.94 4.46 21.98% Multinational 57.52 17.27 3.24 23.76% Family 28.17 39.36 35.25 23.17% Banks 40.83 4.78 1.65 19.41% State 51.85 17.50 0.26 14.06% All 44.34 20.65 10.20 100%

7

Some authors use other variables relating assets or equity market value to assets or equity book value (McConnell and Servaes, 1995). The diference between market and book value proxies growth opportunities' value facing the firm and is supposed to be inversely related to the asset-in-place value. The market-to-book ratio will be used later as a sorting variable in order to test the robustness of the results. Some other variables having been used are the market equity value to total asset ratio (Lasfer, 1995), the market asset value to cash ?ow ratio (Smith and Watts, 1992) or sales' rate of growth (McConnell and Servaes, 1995; La Porta et al., 2000). 8 This ratio was computed by using equity book value. As a robustness test, calculations based on market value were run. Results remain basically unchanged as displayed in Table 1. 9 Perhaps the theoretical justification of this group of variables was not highlighted enough in the first sections of the paper. Now we would like to underline their appropriateness given the remarkable Spanish firm ownership concentration compared with other countries with a more market oriented financial system, more dispersed ownership and less important role for blockholders - such as the UK or the USA (Berglof, 1990; Allen and Gale, 1994; Prowse, 1994; Andre´s and Lo´pez, 1997; Franks and Mayer, 1997).

Financial decisions and growth opportunities
Besides the above mentioned variables, it is usual in this kind of research to include some control vari- ables in order to embody some additional deter- minants of value creation like R&D investment or publicity expenses (Lang and Stulz, 1994; McConnell and Servaes, 1995). Unfortunately, this information is unavailable, so the study has been able just to control for the other two of the most often cited issues: firm size and industry classification. First, LOGMV variable (market value logarithm) represents firm size and, to some extent, it proxies the problems stemming from asymmetric information (Devereux and Schiantarelli, 1990). Second, dummy industry variables were included and more indepth comments about their in?uence can be found in the sensitivity analysis paragraphs. Table 3 provides some information about mean values in the groups the sample was divided into 10 and a test for the diferent mean value hypothesis. As shown, variable means depend heavily on growth opportunities' avalilability and are quite similar, specially in the PER and MB-based classifications. Perhaps it is worth stressing dividend value since a positive relationship between growth opportunities and dividend payout can be found. It is very consis- tent with La Porta et al. (2000) results showing that Spain seems to be one of the few countries supporting the 'substitute model' and high growth companies tend to pay more dividends.

397
cause an omission bias and distort the results. On the other hand, the dynamic dimension of a panel data enhances testing long time adjusting processes and determining the firm value reaction when the explanatory variables change. With regard to the basic model to be estimated, a multivariate regression model has been built includ- ing most of the previously cited variables. It simulta- neously takes into account some issues such as corporate financing, dividend payout and ownership structure. This model can be expressed with the following equation, where i refers to the firm and t to the year (i ¼ 1 . . . 101; t ¼ 1 . . . 5) Qit ¼ ffi þ fi1DTAit þ fi2DIVTAit þ fi3ADJALFAit þ fi4LOGMVit þ fi5C1it þ fi6BAit þ fi6MULTit þ fi7DOMit þ "it The so-specified model was independently tested for each one of the two sub-samples into which the initial sample had been split. 11 The results of the panel data estimation are displayed in Tables 4-6. The estimations were run not only for the basic spec- ification (Tables 4a and 4b) but also the State owned companies were dropped out (Tables 5a and 5b) and firm industry characteristics were introduced (Tables 6a and 6b). The F-test value underlines the existence of an individual efect to the extent that the null hypothesis of individual efect absence is rejected nearly at a 99% confidence level and corroborates the appropriateness of a panel data approach. Furthermore, the Hausman test reveals the importance of the fixed efect component - closely correlated with the remainder explanatory variables - so that the within groups estimation method becomes necessary in order to deal with the constant unobservable heterogeneity.

Methodology As stated before, the sample combines 101 observations with five cross-sections originating a 505observations panel data. Given the aim of the study, the panel data methodology seems to be the most accurate for at least two reasons (Arellano and Bover, 1990; Arellano, 1993). On the one hand, this method allows the control of the so-called unobserv- able constant heterogeneity. It is quite convincing that each one of the firms in the sample has its own specificity - e.g., the way it is run by the managers, the impression it makes to the market, the way it generates growth opportunities, etc. This specificity is diferent from a company to another one and it is almost certain to be kept throughout the study period. A pooling analysis of all the companies without noticing these peculiar characteristics could

IV. Results Results report A general outlook to the basic results shows some interesting issues. For instance, there is a group of explanatory variables coming out significant to an acceptable level. Moreover, the significance of the whole model - both in terms of the R 2 and the adjusted R2 coefcients - is high enough, specially for those companies having more growth opportunities.

10

Although the sorting out criteria will be explained later, it may be interesting to report now the mean value of all the variables for each group. 11 Since regression results could be afected by multicollinearity problems, possible multicollinearity was previously controlled for by running multicollinearity tests and single regressions. The results do not support the existence of multicollinearity.

398

P. de A. Alonso et al.

Table 3. Descriptive statistics Mean values for the groups the sample was divided into and p-value for the test for the diferent mean value hypothesis. (m) stands for market values whereas (b) stands for book values. PER Criteria DTA (m) DTA (b) DBDT C1 (%) C2 (%) C5 (%) ALFA (%) ADJALFA (%) DOM MULT FAM BA STAT LOGMV Q DIVTA (%) Assets Low growth 0.6426 0.5527 0.4481 45.5977 54.8376 62.3188 23.9933 12.3525 0.1548 0.3214 0.2679 0.1726 0.0893 4.1915 0.9131 0.5170 34 057 High growth 0.3984 0.3932 0.3963 46.2818 57.0434 65.2179 20.7860 9.3329 0.2367 0.2308 0.2426 0.2130 0.1065 4.6886 1.1674 1.6650 265 561 p-value 0.000 0.000 0.226 0.403 0.202 0.122 0.130 0.058 0.029 0.031 0.298 0.174 0.298 0.000 0.000 0.000 0.000 Low growth 0.6950 0.5060 0.5046 43.4877 53.0559 59.8302 20.1313 9.2664 0.2262 0.2619 0.2440 0.1190 0.1607 4.3284 0.6970 0.4369 116 491 MB High growth 0.3115 0.4153 0.3211 46.4777 56.8831 65.1569 20.4221 10.1693 0.2249 0.2426 0.2189 0.2308 0.1183 4.7618 1.4674 1.9959 114 955 p-value 0.000 0.002 0.000 0.291 0.160 0.039 0.913 0.622 0.976 0.684 0.586 0.006 0.262 0.000 0.000 0.000 0.948 Low growth 0.5139 0.4713 0.3929 43.3502 53.9117 62.4772 21.8542 11.7999 0.2667 0.2963 0.2667 0.1778 0.0148 4.2102 1.0213 1.0052 109 065 SRGR High growth 0.4947 0.4715 0.3983 46.1947 55.9295 63.6716 20.9316 10.8668 0.2132 0.2426 0.2279 0.2132 0.1250 4.5051 1.0434 1.6045 54 796 p-value 0.551 0.992 0.858 0.357 0.515 0.707 0.729 0.661 0.288 0.303 0.440 0.388 0.000 0.000 0.702 0.037 0.110

Table 4. Value creation determinants conditional on growth opportunities The sorting out criteria was the PER ratio: we divided the whole sample into three groups (each one containing 170 observations) and selected the upper and the lowest third as those firms with more, and less, growth opportunities respectively. The table shows estimated coefcients, t-statistics, p-value and the (adjusted) determination coefcient. Hausman test allows to test fixed versus random efects hypothesis. Hausman test follows a _ 2(8) distribution. The model to be estimated is Q ¼ff þ fi1DTA þ fi2DIVTA þ fi3ADJALFA þ fi4LOGMV þ fi5C1 þ fi6BA þ fi7MULT þ fi8DOM þ " : (a) Absence of growth opportunities Within DTA DIVTA ADJALFA LOGMV C1 BA MULT DOM R2 Adj.-R2 F(33, 128) Hausman test Coefcient 0.632966 0.051433 0.21460E-02 0.227256 À0.39847E-02 À0.367624 0.051162 0.050707 0.624260 0.503906 2.7071 22.263 t-statistic 6.39101 2.35028 0.457513 1.84139 À1.14315 À1.30900 0.283669 0.312308 p-value [0.000] [0.020] [0.648] [0.068] [0.255] [0.193] [0.777] [0.755] (b) Presence of growth opportunities Coefcient À0.309555 À0.69649E-03 0.014829 0.586986 À0.50503E-02 0.479124 0.985452 0.123302 0.837846 0.785906 8.4351 30.427 t-statistic À1.59504 À0.078157 2.80334 8.64295 À2.39484 2.72095 5.37656 1.04082 p-value [0.113] [0.938] [0.006] [0.000] [0.018] [0.007] [0.000] [0.300]

[0.000] [0.004]

[0.000] [0.000]

These results confirm the hypothesis about the in?uence of leverage, dividends and ownership struc- ture on firm value. First, the financial leverage ratio is significant in all the estimations, although its role is quite diferent depending on the existence or the absence of growth opportunities. When firms lack

those profitable projects (Tables 4a, 5a and 6a), the DTA positive sign suggests the debt contribution to firm value creation by disciplining managers. If this is the case, the debt burden reduces the free cash ?ow problem (Jensen, 1986) and prevents managers from wasteful uses from the shareholders' point of view.

Financial decisions and growth opportunities
Table 5. Non-State firms value creation determinants Original regressions are run after dropping State owned and electric companies. (a) Absence of growth opportunities Within DTA DIVTA ADJALFA LOGMV C1 BA MULT DOM R2 Adj.-R2 F(32, 122) Hausman test Coefcient 0.633354 0.052062 0.4227E-2 0.203861 À0.4590E2 À0.277176 0.169111 0.157497 0.636264 0.517006 2.7680 26.123 t-statistic 6.3976 2.3537 0.81398 1.6453 À1.2825 À0.9049 0.7306 0.7709 p-value [0.000] [0.020] [0.417] [0.102] [0.202] [0.367] [0.466] [0.442] (b) Presence of growth opportunities Coefcient À0.384598 0.7206E-4 0.014864 0.565303 À0.57181E-2 0.494302 0.968261 0.119041 0.828131 0.770290 6.327 26.914 t-statistic À1.8626 0.7638E-2 2.6967 7.9572 À2.4519 2.6943 5.0626 0.9642

399

p-value [0.065] [0.994] [0.008] [0.000] [0.016] [0.008] [0.000] [0.337]

[0.000] [0.001]

[0.000] [0.000]

Table 6. Value creation determinants (with industry efects) Original regressions are run including industry dummies. The industries included are: Food and Beverage, Building, Property, Transportation and Communication, Electrical, Chemicals, Metal-mechanical, Mining and Textile and Paper. Automobile and Trade and Retailing were excluded in order to avoid multicolineality. (a) Absence of growth opportunities Random DTA DIVTA ADJALFA LOGMV C1 BA MULT DOM FOOD BUILD PROP TRANS CHEM MET TEXT C2 R Adj.-R2 Hausman test Coefcient 0.616459 0.081163 0.40038E-2 0.172668 0.11536E-3 0.071736 0.047322 0.18178E-2 À0.029371 0.062524 0.325907 0.169121 À0.036876 0.079845 À0.013962 À0.279429 0.503880 0.301297 26.470 t-statistic 6.69826 4.09519 1.43320 2.14686 0.059415 0.464097 0.379349 0.013924 À0.187903 0.439239 1.74166 1.15447 À0.270871 0.420783 À0.111499 À0.715493 p-value [0.000] [0.000] [0.152] [0.032] [0.953] [0.643] [0.704] [0.989] [0.851] [0.660] [0.082] [0.248] [0.786] [0.674] [0.911] [0.474] (b) Presence of growth opportunities Coefcient À0.399162 0.010658 0.018632 À0.514839 À0.3138E-02 0.221984 0.540189 0.199290 0.211593 0.077113 À0.032228 À0.410314 0.025676 0.173098 À0.116519 À1.33008 0.770277 0.676473 33.642 t-statistic À2.21145 1.21879 4.63449 8.69286 À1.81514 1.55699 4.22334 1.99969 0.840284 0.348929 À0.113552 À1.37640 0.082483 0.720033 À0.437270 À4.59069 p-value [0.027] [0.223] [0.000] [0.000] [0.070] [0.119] [0.000] [0.046] [0.401] [0.727] [0.910] [0.169] [0.934] [0.472] [0.662] [0.000]

[0.001]

[0.000]

On the contrary, DTA coefcient becomes negative in the estimation for the most highly priced compa- nies (Tables 4b, 5b and 6b), emphasizing the negative impact that debt can have on firm value when firms face growth opportunities, as suggested by the under- investment hypothesis (Myers, 1977). In comparison with the values in Tables 4a-6a, the confidence level

is slightly lower - although fairly acceptable - and the absolute value of the coefcient is around half of the value achieved in the absence of growth opportunities. This last result could be understood as the more important role that debt plays in the overinvestment framework related to the underinvestment one.

400
Second, dividend policy (DIVTA) also takes part in the determination of firm value. Since dividends can be conceived as a way to reduce manager discretion- ality, a diferent sign of the variable is forecast depending on the growth opportunities' availability. The results confirm this dual behaviour, although the confidence level is notably diferent. When companies do not have growth opportunities DIVTA is sig- nificant and positively related to firm value (see Tables 4a-6a), whereas if companies face profitable investment projects dividends exhibit a scarce - and sometimes negative - impact (see Tables 4b-6b). This evidence comes to confirm the hypothesis concerning the disciplinary role of dividends, while the pertinence of earnings retention to fund valuable growth opportunities has no empirical support in the present study. The significance of dividend policy is compatible with the leverage in?uence since both of them seem to have remarkable impact on firm value. The underlying intuition is that debt and dividends are complementary mechanisms to cope with managers' discretionality rather than alternative ways of moni- toring and control: a company having a high leverage ratio does not imply the rejection of dividend policy as a disciplinary mechanism and vice versa. Third, as far as the ownership structure variables are concerned, it is worth noticing the negative impact the ownership concentration (C1) has in both sub-samples, although it only comes out sig- nificant when firms have growth opportunities. In the authors'opinion, this result is consistent with the previous set of results and demonstrates again the existence of some agency problems inside the companies. In the face of growth opportunities, a majority control seems to be disadvantageous, the need of specialized managers can be inferred, and the ownership-control separation arises not to be so harmful - for shareholders' wealth - as usually thought (Burkart et al., 1997). Furthermore, the existence of blockholders in the ownership structure of companies with growth opportunities could lead, to some extent, to waste of these opportunities as Carlin and Mayer (1998) have suggested. Regarding the proportion of the shares the directors own (ADJALFA), this variable behaves as predicted, although only partially. In spite of the fact that the coefcient, as forecast, is always positive, it supports only partially the proposed hypothesis because the variable is significant only in the sub- sample of companies with growth opportunities. Obviously, its positive correlation with firm value denotes the convergence of directors and share- holders interests. The more prominent efect in the group of firms with growth opportunities could be

P. de A. Alonso et al.
reasonably explained on the basis of the signalling theory since companies with the best growth opportunities set will try the market to notice their opportunities in order to overcome the information asymmetries, so that the market reacts positively in face of the signal. The ownership structure efect is completed by introducing a number of dummy variables con- cerning the nature of the main shareholder. Some caveats are required to rightly analyse these results because the lack of hypothesis about their possible in?uence - mainly as a result of the lack of an appro- priate theoretical framework - prevents one from drawing concluding evidence. In general, it is found that, for companies with growth opportunities firm value is positively correlated with the main shareholder being a bank or a multinational firm, consistently with some recent research (Khanna and Palepu, 1999). However, one is not able to assert if this kind of owner positively afects firm value or, on the contrary, these main shareholders select the companies with the best growth opportunities. In this last case, ownership structure would not be the cause but the consequence of firm valuation. The last comments focus on the control variables. Company size (LOGMV) was no object of theoretical prediction because this feature is out of the initial purpose. Nevertheless, that variable was included in order to control for the size efect. Firm size comes out to be clearly significant and positively related to firm value in all the estimations. There is a wide range of possible explanations, but most of them rely on the idea of information asymmetries or, in other words, the size of the company as a syno- nym of being better known in capital markets and, hence, of better reputation. Finally, neither individu- ally, nor together, were the industry dummies found to have any significant efect in each one of the sub-samples. It should be noted that this set of variables makes sense only in the random efects model (Table 6) since industry variables are constant throughout the period and hence their efect is removed by estimating the within groups method - the most suitable method as the Hausman test indicates. Sensitivity analysis One of the study's concerns is to know whether the results that have been obtained are contingent upon the specification of the model. In order to assess the robustness of the results to alternative specifications and variable measurements a sensitivity analysis is added consisting of four diferent tests: a change in sample composition, an alternative identification of

Financial decisions and growth opportunities
growth opportunities, the incorporation of industry specific features and a non-linear relationship with directors' ownership. Regarding sample composition, one wonders if the inclusion of State owned companies could bias the results. State owned companies may be afected by a number of very specific and not easy to generalize circumstances like non-profit but public service aim, potential monopoly situations, specially regulated industries, State subsidies policies, etc.12 Based on these particular features, the regressions have been run after dropping the 16 State companies and utilities.13 The results for non-State firms, reported in Table 5, shows no noticeable change respective of the whole sample estimation with the only excep- tion of leverage increasing its signification and becoming significant to a 93.5% confidence level. As previously stated, growth opportunities identification is a key aspect in the study and this is why an analysis of the in?uence of the way one mea - sures these growth opportunities is considered vital. In addition to the PER ratio, the same regressions have been run using the market-to-book equity ratio and sales rate of growth. Regarding the first variable, it is a meaningful proxy to market expectations about firm projects profitability: those companies with more valuable opportunities should have higher equity market value relative to book value - and hence, should display higher ratios than those lacking of profitable projects. Therefore, the sample has been ranked on the basis of this ratio and the upper and the lowest third taken as those companies with the widest and narrowest set of profitable growth opportunities respectively. Table 7 reports the results, which are greatly consistent with those previously obtained both in the presence and in the absence of growth opportunities. Debt not only keeps on being highly significant when firms lack valuable projects but also becomes signifi- cant at 99% in the presence of growth opportunities. Also consistent with PER estimations, dividend payment significantly contributes to value creation in the absence of growth opportunities, although there is a slight reduction in the confidence level - none the less, this level remains higher than 90%. Similarly, ownership and control variables remain basically with the same in?uence that was previously

401
detected, with the only change in the nature of the main shareholder. When firms have some profitable growth opportunities, ownership concentration and directors' ownership percentage carry on having significant impact on firm value whereas they do not have any significant efect in their absence. As far as other ownership variables are concerned, the conclusion remains unafected: the positive contribu- tion of companies - either banks or multinational firms - to managers' monitoring and agency con?icts resolution. Firm size also seems to have a positive role in firm value creation. As regards to the model explanatory power (R2 and adjusted-R2 coefcients), although it shows some changes (it increases anddecreases in the absence or in the presence of growth opportunities respectively) it seems to remain quite acceptable. Finally, it should be noted that the Hausman test reveals the lack of correlation between the fixed individual efects term and the set of the independent variables, suggesting the generalized least squares regression as the most efcient method rather than within groups estimation (Arellano, 1990). The market-to-book ratio regressions were also run after excluding State companies. The results (Table 8) require no further comments and are consistent with previous ones. In spite of relying on the past as a proxy for the future, sales rate of growth has also been used to proxy growth opportunities (McConnell and Servaes, 1995; La Porta et al., 2000). This is why the sample has been ranked by the sales rate of growth (SRGR) and the previous regressions run in each one of the two usual groups. Results are presented in Tables 9-10 and show how debt and dividends can mitigate agency problems in lowgrowth firms while they have no significant in?uence when profitable projects are available. Compared with the PER or MB models, the only remarkable diference is the impact of C1 on firm value. This result highlights the dual role for ownership concen- tration, so that it enhances value creation if compa- nies have no profitable projects but it may destroy value when companies face growth opportunities. There are some other variables related to the main shareholder nature (MULT, BA and DOM) coming out as partially significant determinants of value and underlining the necessity to control for

12

State incumbency in firm ownership has dramatically changed since 1996 when Spanish Government undertook a privatization programme. However, since the sample covered the 1991-1995 period it has been considered pertinent to test the possible bias due to State companies inclusion. From the point of view of the present study, the privatization process might make this caveat no longer necessary if the sample was extended to more recent years. 13 Utilities have been excluded because of the many aspects they have in common with State firms in Spain.

402

P. de A. Alonso et al.
Table 7. Value creation determinants conditional on growth opportunities (with MB) Original regressions are run after changing the sorting out criteria. The sorting out criteria was the valuation ratio (MB): the whole sample was divided into three groups (each one containing 170 observa- tions) and the upper and the lowest third selected as those firms with more, and less, growth opportunities respectively. The leverage extreme values have been dropped out. The table shows estimated coefcients, t-statistics, p-value and the (adjusted) determination coefcient Hausman test allows to test fixed vs. random efects hypothesis. The model to be estimated is Q ¼ff þ fi1DTA þ fi2DIVTA þ fi3ADJALFA þ fi4LOGMV þ fi5C1 þ fi6BA þ fi7MULT þ fi8DOM þ ": (a) Absence of growth opportunities Within DTA DIVTA ADJALFA LOGMV C1 BA MULT DOM C2 R Adj.-R2 F(60,94) Hausman test _2(8) Coefcient 0.592585 0.015215 0.4723E-3 0.038523 0.6656E-4 0.032170 0.022328 0.036772 0.200101 0.766351 0.600147 8.8342 t-statistic 14.9779 1.7125 0.6841 2.8726 0.1907 1.0797 1.0031 1.7465 3.3343 p-value [0.000] [0.087] [0.494] [0.004] [0.849] [0.280] [0.316] [0.081] [0.000] (b) Presence of growth opportunities Coefcient À1.013400 0.3210E-2 0.012034 1.124610 À0.6744E-2 0.580646 0.836602 0.139186 0.501003 0.140026 5.6029 51.000 t-statistic À2.9180 0.2785 1.9279 6.7811 À2.7192 3.3007 4.9720 0.7770 p-value [0.004] [0.781] [0.057] [0.000] [0.008] [0.001] [0.000] [0.493]

[0.356]

[0.000] [0.000]

Table 8. Non-State firms value creation determinants (with MB) Original regressions are run after dropping State owned and electric companies. (a) Absence of growth opportunities Within DTA DIVTA ADJALFA LOGMV C1 BA MULT DOM C2 R Adj.-R2 F(54, 78) Hausman test _2(8) Coefcient 0.598570 0.016750 0.4014E-3 0.036282 0.1878E-3 0.028478 0.015587 0.031768 0.205584 0.808239 0.664418 9.4739 t-statistic 14.1491 1.7159 0.5154 2.1903 0.4624 0.8003 0.5309 1.1493 3.1227 p-value [0.000] [0.086] [0.606] [0.028] [0.644] [0.424] [0.595] [0.250] [0.000] (b) Presence of growth opportunities Coefcient À1.195800 0.6563E-2 0.7984E-2 0.951189 À0.7792E-2 0.546805 0.831025 0.074640 0.841075 0.714750 5.8922 39.265 t-statistic À3.2427 0.5561 1.2512 5.4426 À2.9013 3.1265 4.9678 0.4143 p-value [0.000] [0.580] [0.215] [0.000] [0.005] [0.002] [0.000] [0.680]

[0.303]

[0.000] [0.000]

ownership structure. R2 and adjusted-R2 coefcients exhibit similar values to those of previous regressions ranging from 0.7 to 0.8. This sensitivity analysis section is continued by intention to control for industry heterogeneity in case diferent industries were in diferent business cycle positions, faced diferent regulatory frameworks

and, in turn, had very diferent growth opportunities. If this was the case, one could have found spurious relationships since two firms belonging to quite diferent industries are not comparable on the basis of their growth opportunities because of the very industry-specific content of these opportunities. To control for industry heterogeneity a set of dummy

Financial decisions and growth opportunities

403

Table 9. Value creation determinants conditional on growth opportunities (with SRGR) Original regressions are run after changing the sorting out criteria. The sorting out criteria was the sales rate of growth (SRGR): the whole sample was divided into three groups (each one containing 170 observations) and the upper and the lowest third selected as those firms with more and less growth opportunities respectively. The leverage extreme values have been dropped. The table shows estimated coefcients, t-statistics, p-value and the (adjusted) determination coefcient Hausman test allows to test fixed vs. random efects hypothesis. The model to be estimated is

Q ¼ff þ fi1DTA þ fi2DIVTA þ fi3ADJALFA þ fi4LOGMV þ fi5C1 þ fi6BA þ fi7MULT þ fi8DOM þ ":
(a) Absence of growth opportunities Within DTA DIVTA ADJALFA LOGMV C1 BA MULT DOM R2 Adj.-R2 F(33,128) Hausman test _2(8) Coefcient 0.655851 0.024091 1.65E-03 0.416804 5.17E-03 À0.052771 0.335069 0.049459 0.804116 0.741371 7.89949 18.153 t-statistic 7.11206 2.21566 0.333207 3.53662 1.84671 À0.27141 2.60505 0.461061 p-value [0.000] [0.028] [0.740] [0.001] [0.067] [0.787] [0.010] [0.646] (b) Presence of growth opportunities Coefcient 0.220634 À9.14E-03 0.01427 0.3875 À8.67E-03 0.263386 0.586934 0.410526 0.773285 0.700665 6.3011 23.436 t-statistic 1.4538 À0.494296 4.82812 6.43566 À5.07287 1.66016 4.23683 3.33986 p-value [0.148] [0.622] [0.000] [0.000] [0.000] [0.099] [0.000] [0.001]

[0.0000] [0.0201]

[0.0000] [0.0028]

Table 10. Non-State firms value creation determinants (with SRGR) Original regressions are run after dropping out State owned and electric companies. (a) Absence of growth opportunities Within DTA DIVTA ADJALFA LOGMV C1 BA MULT DOM R2 Adj.-R2 F(30, 110) Hausman test _2(8) Coefcient 0.65587 0.022499 2.54E-03 0.391722 4.89E-03 4.65E-03 0.402445 0.093679 0.785755 0.711743 6.0931 18.546 t-statistic 6.66766 1.90616 0.465187 3.05769 1.49153 0.021578 2.6339 0.69623 p-value [0.000] [0.059] [0.643] [0.003] [0.139] [0.983] [0.010] [0.488] (b) Presence of growth opportunities Coefcient 0.139059 À0.014207 0.013991 0.353328 À8.63E-03 0.26107 0.561004 0.394524 0.788424 0.719115 6.5274 20.076 t-statistic 0.981197 À0.830741 5.23 6.41619 À5.57107 1.82137 4.4755 3.54139 p-value [0.329] [0.408] [0.000] [0.000] [0.000] [0.071] [0.000] [0.001]

[0.0000] [0.0175]

[0.0000] [0.0101]

variables have been added to the independent vari- ables concerning the industry the firm belongs to (Table 6). For the sake of brevity there is no comment on the results since, broadly speaking, they agree with those previously exposed, specially as far as debt and dividend in?uence is concerned. The last sensitivity analysis has dealt with the inclusion of a quadratic term for directors' ownership (ADJALFA2). As proved by some authors (Morck

et al., 1988), a non-linear relationship between firm value and directors' ownership can be found because of a trade-of between convergence and entrenchment motivations. This is why a quadratic term or even a piecewise regression would be worthwhile. Table 11 shows the regression results when ADJALFA2 is incorporated. It comes out significant only for the high growth firms and, as forecast by theory, it has a negative impact on firm value, just the

404

P. de A. Alonso et al.

Table 11. Value creation determinants conditional on growth opportunities (with ADJALFA2) The sorting out criteria was the PER ratio: the whole sample was divided into three groups (each one containing 170 observations) and the upper and the lowest third selected as those firms with more, and less, growth opportunities respectively. The table shows estimated coefcients, t-statistics, p-value and the (adjusted) determination coefcient. Hausman test allows testing fixed versus random efects hypothesis. Hausman test follows a _2(8) distribution. The model to be estimated is Q ¼ff þ fi1DTA þ fi2DIVTA þ fi3ADJALFA þ fi4ADJALFA2 þ fi5LOGVM þ fi6C1 þ fi7BA þ fi8MULT þ fi9DOM þ ": (a) Absence of growth opportunities Within DTA DIVTA ADJALFA ADJALFA2 LOGMV C1 BA MULT DOM R2 Adj.-R2 F(33, 127) Hausman test Coefcient 0.631797 0.050478 À0.00974437 0.00020393 0.197797 À0.00495088 À0.432706 0.02347 0.044833 0.627913 0.504861 2.5524 24.327 t-statistic 6.38501 2.30713 À0.83749 1.11662 1.56873 À1.37973 À1.50999 0.12904 0.276252 p-value [0.000] [0.023] [0.404] [0.266] [0.119] [0.170] [0.134] [0.898] [0.783] (b) Presence of growth opportunities Coefcient À0.314459 0.00474149 0.057373 À0.00079805 0.586879 À0.0027791 0.546871 1.01012 0.242047 0.847369 0.796893 8.6003 30.32 t-statistic À1.66349 0.533224 3.5931 À2.81497 8.87203 À1.25933 3.15764 5.65145 1.97017 p-value [0.099] [0.595] [0.000] [0.006] [0.000] [0.210] [0.002] [0.000] [0.051]

[0.0001] [0.0039]

[0.0000] [0.0004]

opposite of ADJALFA. In any case, the remaining variables present no remarkable change, nor do R 2 and adjusted-R2 coefcients.

V. Concluding Remarks The classical debate about the relevance of financial decisions on firm value is notably broadened by introducing the set of growth opportunities at the reach of the firm. Then, traditional frameworks fade and debt, dividend payout and ownership and control structure arise as diferent factors afecting thoseopportunities' utilization. In essence, these mechan- isms try to give managers the incentive to efciently use firm cash ?ows and to impede wasteful uses. On the one hand, following the underinvestment incentive (Myers, 1977), very highly leveraged firmscould forego profitable investment projects. On the other hand, as stated by Jensen (1986), debt could lead managers to reject unprofitable projects that may increase managers' utility. Similarly, dividend payments, as long as it may reduce free cash ?ows, acts in the same way, specially when companies do not face growth opportunities and, hence, more intense the overinvestment problem can become. The relevance of financing and dividend decisions is reinforced by including some ownership structure variables. In fact, there is a growing stream in recent

literature focusing on the relationship between some ownership and control structure issues - such as ownership concentration or the monitoring role of the board of directors - and firm growth opportunities. This theoretical framework has been applied to a sample of large Spanish companies publicly traded in capital markets for the 1991-1995 period. The results confirm most of the predicted hypotheses concerning the role of debt, dividends and ownership and control structure. First, leverage assumes a double and active role: it helps to create value by disciplining managers in those companies with no or very scarce growth opportunities, while it has a negative efect in those firms with the best opportunities due to the propen- sity to forgo profitable projects. Second, dividend in?uence basically follows the same pattern, and a positive and significant correlation between dividend payment and firm value has been found in the absence of growth opportunities - the situation when a too high earnings retention could more likely originate inefcient investments. An outstanding feature of the results is the fact that debt and dividends do not seem to be mutually excluding mechanism: those companies requiring a more in-depth monitoring and control use simultaneously debt and dividends as complementary ways to avoid possible free cash ?ow abuses. Regarding ownership structure, we have found an unequivocal linkage to firm value, although we have

Financial decisions and growth opportunities
not always been able to identify the underlying causality relationship. We have detected some in?uence of managers ownership and ownership concentration on firm value, along with some features of the main shareholder nature - being a bank or a multinational firm- positively related to value creation. These results are largely consistent with a test for a non- linear relationship between directors ownership and firm value as suggested by the convergence and entrenchment hypothesis. The explanation of these results relies on an incentive and monitoring approach, emphasizing the usefulness of ownership as a way to give managers incentive and to give rise to an efcient control. To sum up, it is thought that, broadly speaking, the study confirms the already existing intuitions about the possible relationship between financing and divi- dend decisions, contractual structure, growth oppor- tunities and firm market valuation. The results achieved are consistent with those obtained by a number of authors from other countries. Some future research directions can be pointed at as the extension of the sample to an international basis in order to elucidate if country-specific factors such as the finan- cial system design or the firm-bank relationship network can dramatically modify the conclusions achieved. The authors would also like to examine in greater depth the causality relationship among some of the most significant variables, and incorporate a more detailed industry classification.

405
Andre´s, P., Azofra, V. and Rodr?´guez, J. A. (2000) Endeudamiento, oportunidades de crecimiento y estructura contractual. Un contraste emp?´rico para el caso espan˜ol, Investigaciones Econo´micas, 24(3), 641-79. Arellano, M. (1993) Introduccio´n al ana´lisis econome´trico con datos de panel, La industria y el comportamiento de las empresas espan˜olas. Ensayos en homenaje a Gonzalo Mato, Alianza Editorial, Madrid, pp. 23-47. Arellano, M. and Bover, O. (1990) La econometr?´a de datos de panel, Investigaciones Econo´micas (Segunda e´poca), 14(1), 3-45. Barclay, M. J. and Smith, C. W. (1996) On financial architecture: leverage, maturity and priority, Journal of Applied Corporate Finance, 8(4) 4-17. Becht, M. and Ro¨ell, A. (1999) Blockholdings in Europe, European Economic Review, 43(4-6), 1049-56. Berger, P. G. and Ofek, E. (1995) Diversification's efect on the firm value, Journal of Financial Economics, 37, 3965. Berglo¨f, E. (1990) Corporate Control and Capital Structure. Essays on Property Rights and Financial Contracts, Institute of International Business, Stockholm. Bergstro¨m, C. and Rydqvist, K. (1990) The determinants of corporate ownership, Journal of Banking and Finance, 14, 237-53. Bhattacharya, S. (1979) Imperfect information, dividend policy, and 'the bird in the hand' fallacy, Bell Journal of Economics &Management Science, 10(1), 259-70. Bhattacharya, S. and Thakor, A.V. (1993) Contemporary banking theory, Journal of Financial Intermediation, 3(1), 250. Burkart, M., Gromb, D. and Panunzi, F. (1997) Large shareholders monitoring, and the value of the firm, Quarterly Journal of Economics, 112, 693-728. Campbell, T. S. (1979) Optimal investment financing decisions and the value of confidentiality, Journal of Financial and Quantitative Analysis, 14(5), 913-24. Carlin, W. and Mayer, C. (1998) Finance, investment and growth, Mimeo, Sa?¨d Business School, University of Oxford. Chung, K. and Charoenwong, C. (1991) Investment options, assets in place, and the risk of stocks, Financial Management, 20, 21-33. Demsetz, H. (1983) The structure of ownership and the theory of the firm, Journal of Law and Economics, 26(2), 375-90. Demsetz, H. and Lehn, K. (1985) The structure of corporate ownership: causes and consequences, Journal of Political Economy, 93(6), 1155-77. Devereux, M. and Schiantarelli, F. (1990) Investment, financial factors, and cash ?ow: evidence from UK panel data, in Asymmetric Information, Corporate Finance and Investment (Ed.) R. G. Hubbard, The University of Chicago Press, Chicago, pp. 279-306. Fama, E. F. and Jensen, M. C. (1983) Separation of ownership and control, Journal of Law and Economics, 26(2), 301-25. Fazzari, S. M., Hubbard, R. G. and Petersen, B. C. (1988) Financing constraints and corporate investment, Brooking Papers on Economic Activity, 1, 141-95. Franks, J. and Mayer, C. (1997) Corporate ownership and control in the UK, Germany and France, Journal of Applied Corporate Finance, 9(4), 30-45.

Acknowledgements The authors are grateful to M. Ferna´ndez, S. Go´mez and M. Ewing, seminar participants in the IV Work- shop in Finance (Segovia), 2000 European Financial Management Association Meeting (Athens), 10th Asociacio´n Cient?´fica de Econom?´a y Direccio´n de la Empresa Conference (Oviedo) for their comments on earlier versions of the paper. This research has received financial support from the Spanish Direccio´n General de Ensen˜anza Superior e Investigacio´n Cient?´fica (PB97-0594) and from the Junta de Castilla y Leo´n (VAOS204). All the remain- ing errors are the authors' own responsibility.

References
Allen, F. and Gale, D. (1994) A welfare comparison of the German and US financial systems, Working Paper, CEPRBBV. Andre´s, P. and Lo´pez, F. J. (1997) Financial system models, corporate governance and capital investment in OECD countries: some stylized facts, European Investment Bank Papers, 2(2), 69-96.

406
Gonza´lez, F. (1995) La reaccio´n de los precios de las acciones ante anuncios de dividendos: la evidencia emp?´rica en el mercado espan˜ol de valores, Investigaciones Econo´micas, 19(2), 249-68. Harris, M. and Raviv, A. (1991) The theory of capital structure, Journal of Finance, 46(1), 297-355. Jensen, M. C. (1986) Agency costs of free cash ?ow, corporate finance, and takeovers, American Economic Review, 76, 323-9. Jensen, M. C. (1993) The modern industrial revolution, exit, and the failure of internal control systems, Journal of Finance, 48(3), 831-80. Jensen, M. C. and Meckling, W. (1976) Theory of the firm: managerial behavior, agency costs and owner- ship structure, Journal of Financial Economics, 3(4), 305-60. Khanna, T. and Palepu, K. (1999) Emerging market business groups, foreign investors, and corporate control, NBER Working paper, No. 6955 La Porta, R., Lo´pez de Silanes, F. and Shleifer, A. (1998) Corporate ownership around the world, Journal of Finance, 54(2), 471-517. La Porta, R., Lo´pez de Silanes, F. and Shleifer, A. (1999) Corporate ownership around the world, Journal of Finance, 54(2), 471-517. La Porta, R., Lo´pez de Silanes, F. and Shleifer, A. (2000) Agency problems and dividend policies around the world, Journal of Finance, 56(1), 1-33. Lang, L. and Litzenberger, R. H. (1989) Dividend announcements, cash ?ow signaling vs. free cash ?ow hypothesis, Journal of Financial Economics, 24, 181-91. Lang, L. and Stulz, R. M. (1994) Tobin's q, corporate diversification, and firm performance, Journal of Political Economy, 102(6), 1248-80. Lang, L., Ofek, E. and Stulz, R. M. (1996) Leverage, investment, and firm growth, Journal of Financial Economics, 40, 3-29. Lasfer, M. A. (1995) Agency costs, taxes, and debt: the UK evidence, European Financial Management, 1(3), 265-85. Leland, H. E. and Pyle, D. H. (1977) Informational asymmetries, financial structure and financial intermediation, Journal of Finance, 32(2), 371-87.

P. de A. Alonso et al.
McConnell, J. J. and Servaes, H. (1990) Additional evidence on equity ownership and corporate value, Journal of Financial Economics, 26, 595-612. McConnell, J. J. and Servaes, H. (1995) Equity ownership and the two faces of debt, Journal of Financial Economics, 39, 131-57. Miller, M. H. and Modigliani, F. (1961) Dividend policy, growth and the valuation of shares, Journal of Business, 34(4), 411-33. Miller, M. H. and Rock, K. (1985) Dividend policy under asymmetric information, Journal of Finance, 40, 1031-51. Modigliani, F. and Miller, M. H. (1958) The cost of capital, corporation finance and the theory of investment, American Economic Review, 68(3), 261-97. Morck, R., Schleifer, A. and Visnhy, R. N. (1988) Management ownership and market valuation, Journal of Financial Economics, 20(1/2), 293-315. Myers, S. C. (1977) Determinants of corporate borrowing, Journal of Financial Economics, 5(2), 147-75. Prowse, S. (1994) Corporate governance in an international perspective: a survey of corporate control mechanism among large firms in the United States, the United Kingdom, Japan and Germany, BIS Economic Papers, No. 41, July. Shleifer, A. and Vishny, R. (1986) Large shareholders and corporate control, Journal of Political Economy, 94(3), 461-88. Smith, C. W. and Warner, J. B. (1979) On financial contracting: an analysis of bond covenants, Journal of Financial of Economics, 7(2), 117-61. Smith, C. W. and Watts, R. (1992) The investment opportunity set and corporate financing, dividend, and compensation policies, Journal of Financial Economics, 32, 263-92. Stiglitz, J. E. (1985) Credit markets and the control of capital, Journal of Money, Credit and Banking, 17(2), 133-52. Stulz, R. (1988) Managerial control of voting rights: financial policies and the market for corporate control, Journal of Financial Economics, 20, 25-54. Stulz, R. (1990) Managerial discretion and optimal financing policies, Journal of Financial Economics, 26, 3-27.

Financial decisions and growth opportunities
Appendix: Variables Glossary
Abbreviations: equity market value (EMV ); equity book value (EBV ); total debt (D); dividend payment (DIV ); total assets (TA); net income (NI ); Sales (S). Abbreviation DTA DIVTA C1 ALFA ADJALFA DOM MULT FAM BA STAT LOGMV PER MB SRGR D/(D þ EBV) DIV/(D þ EBV) Main shareholder participation (%) Directors ownership participation (%) Directors ownership participation (%) (¼1 for domestic companies) (¼1 for multinational companies) (¼1 for families and individuals) (¼1 for banks) (¼1 for State owned) LOG(EMV þ D) (EMV þ D)/(EBV þ D) EMV/NI EMV/EBV (St À StÀ1)/StÀ1 Definition Total debt/Total asset (book value) Dividends/Total asset (%) Ownership concentration. Directors' ownership participation (%) Directors' ownership participation (only natural people) Main shareholder nature Main shareholder nature Main shareholder nature Main shareholder nature Main shareholder nature Size proxy Financial q. Value creation Price-earning ratio. Growth opportunities proxy Valuation ratio. Growth opportunities proxy Sales rate of growth. Growth opportunities proxy Mean 0.4576 0.012231 44.346 20.658 10.201 0.2198 0.2376 0.2317 0.1941 0.1406 4.5206 1.0360 33.095 1.0556 0.7054

407

Std. dev. 0.2529 0.22586 25.832 25.313 18.050 0.4145 0.4260 0.4223 0.3959 0.3479 0.7258 Q 0.4556 43.861 0.8258 4.5594



doc_444424776.docx