Determinants of Dividend Payout Policy: Regulated Versus Unregulated Firms by Atul K. Saxena

Atul K. Saxena saxena_ak@Mercer.EDU is an Associate Professor of Finance at Mercer University.

I. INTRODUCTION

A number of researchers have provided insights, theoretical as well as empirical, into the dividend policy puzzle. However, the issue as to why firms pay dividends is as yet unresolved. Several rationales for a corporate dividend policy have been proposed in the literature, but there is no unanimity among researchers. Everyone, however, agrees that the issue is important, as dividend payment is one of the most commonly observed phenomenon in corporations worldwide.

The issue of dividend policy is important for several reasons. First, researchers have found that a firm uses dividends as a mechanism for financial signaling to the outsiders regarding the stability and growth prospects of the firm. Secondly, dividends play an important role in a firm's capital structure. Yet another set of studies have established the relationship between firm dividend and investment decisions. According to the "residual dividend" theory, a firm will pay dividends only if it does not have profitable investment opportunities, i.e., positive net present value projects.

Further, a firm's stock price is affected, among other things, by the dividend pattern. Firms usually do not like to reduce or eliminate dividend payments [Woolridge and Ghosh, 1988 and 1991], hence, they make announcements of dividend initiation or increases only when they are confident of keeping up with their good performance. Moreover, because the success of a financial manager is tied to the maximization of shareholder wealth (and firm value), hence s/he must understand the dynamics of dividend policy. Indeed, the market value of a firm is dependent upon its stock price. One of the most popular models for stock valuation (the dividends discounting model or DDM) relies upon the assumption that the firm will pay dividends until eternity.

Despite the rich literature on the overall issue of dividend policy, most studies exclude regulated firms from their analyses. The common explanation for excluding regulated firms is that regulators, directly or indirectly, dictate how much dividend the firm can pay. This paper, on the other hand, uses several financial variables to explain the possible differences in the dividend policy of both regulated and unregulated firms. The results indicate that the reasoning proposed by the literature on unregulated firms is correct. Indeed, the signs of parameter estimates correspond with earlier studies.

The study goes a step further, however, in studying regulated firms. While comparing regulated and unregulated firms, the study reveals some interesting results. On average, a regulated firm is less risky; has a lower growth rate; has much fewer insiders holding its common stock; and has fewer investment opportunities, but pays a higher percentage in dividends. Possible explanations for this are provided in this paper. While the issue of dividend policy is far more pervasive, this study tries to address the determinants of dividend payout for firms. Specifically, it is concerned with addressing what factors determine the dividend payout rate. Hence, for the purpose of this paper, a firm's dividend policy is proxied by its dividend payout rate, which is defined as the ratio of dividends per share and earnings per share.

The remainder of the paper is organized as follows: Section II provides a brief review of the relevant literature. Section III presents the description of the data and the empirical methods used. Section IV contains the results, and Section V concludes the paper with a summary.

The theory of industrial organization tells us that there can be three kinds of regulation. Price regulation, geographic market regulation, and product market regulation. It is hypothesized that regulation impacts dividend policy for the following reasons:

Most studies on dividend policy, however, either exclude regulated firms, or analyze them separately.

Miller and Modigliani [1961] view dividend payment as irrelevant. According to them, the investor is indifferent between dividend payment and capital gains. Black [1976] poses the question again, "Why do corporations pay dividends?" In addition, he poses a second question, "Why do investors pay attention to dividends?" Although, the answers to these questions may appear obvious, he concludes that they are not. The harder we try to explain the phenomenon, the more it seems like a puzzle, with pieces that just do not fit together. After over two decades since Black's paper, the dividend puzzle persists.

Some researchers emphasize the informational content of dividends. Miller and Rock [1985], for instance, develop a model in which dividend announcement effects emerge from the asymmetry of information between owners and managers. The dividend announcement provides shareholders and the marketplace the missing piece of information about current earnings upon which their estimation of  the firm's future (expected) earnings is based. The latter, of course, determines the current market value of the firm. In this respect, we can clearly see the role played by dividends. The dividend announcement provides the missing piece of information and allows the market to establish the firm's current earnings. These earnings are then useed in predicting future earnings. John and Williams [1985] construct an alternative signaling model in which the source of the dividend information is liquidity driven.

There are other factors influencing a firm's dividend policy. For example, some studies suggest that dividend policy plays an important role in determining firm capital structure and agency costs. Since Jenson and Meckling [1976], many studies have provided arguments that link agency costs with the other financial activities of a firm. Easterbrook [1984] says that firms pay out dividends in order to reduce agency costs. Dividend payout keeps firms in the capital market, where monitoring of managers is available at lower cost. If a firm has free cash flows [Jensen (1986)], it is better off sharing them with stockholders as dividend payout (or retiring the firm's debt) in order to reduce the possibility of these funds being wasted on unprofitable (negative net present value) projects.

Crutchley and Hansen [1989] examine the relationship between ownership, dividend policy, and leverage and conclude that managers make financial policy tradeoffs to control agency costs in an efficient manner. More recently, researchers have attempted to establish the link between firm dividend policy and investment decisions. Smith and Watts [1992] investigated the relations among executive compensation, corporate financing, and dividend policies. They conclude that a firm's dividend policy is affected by its other corporate policy choices. In addition, Jensen, Solberg, and Zorn [1992] linked the interaction between financial policies (dividend payout and leverage) and insider ownership to informational asymmetries between insiders and external investors. They employed a simultaneous system of equations and found that corporate financial decisions and insider ownership are interdependent. Despite this rich literature, most prior work implicitly recognizes differences in determinants of financial decisions between regulated and unregulated firms by excluding regulated firms from the analysis.

Rozeff [1982] was among the first to explicitly recognize the role of insiders as one of monitoring the managers. He finds that dividend policy for unregulated firms is negatively related to its level of insider holdings. One interpretation of his result is that firms with higher levels of insider holdings have less need to signal firm value through dividends than comparable firms with lower levels of insider holdings. Additionally, in the context of the investment and financing decision, Myers and Majluf [1984] showed that the level of insider holdings is itself a signal of firm value.

In a study of electric utilities, Hansen, Kumar, and Shome [1994] focused on the role that dividends play in the monitoring process to reduce equity agency costs. Hansen et al. focusd on electric utilities since they do not seem to fit current dividend theory explanations in the literature. They act differently, perhaps because they are subject to regulatory oversight and insulated from most market disciplines like takeovers. Their paper concludes that the use of higher payout raises the likelihood of monitoring by both management and the regulatory authority. If the regulator sets the rate of return to shareholders (dividend yield) below that required by market, then assuming efficient markets, the marginal investors will drop out. This lowering of the demand for the company's stock will adversely affect its price reflecting greater difficulty in raising equity funds. Moreover, the associated costs (e.g., transactions and opportunity costs) will go up. Therefore, even if one assumes that this does not affect the costs of other sources of financing, the increased cost of equity financing will result in a higher overall cost of capital for the firm.

Moyer, Rao, and Tripathy [1992] suggested that regulated firms use dividends as a means of subjecting the utility and the regulatory rate commission to market discipline, in keeping with the Smith [1986] hypothesis. Smith [1986] argues that by subjecting the regulatory commission to capital market discipline as the utility raises new capital, the utility can ensure more favorable rate adjustments. Moyer et al. also found that the dividend policies for these firms respond to changes in policies adopted by regulatory commissions. In a related article, Moyer, Chatfield, and Sisneros [1989] found that security analysts' monitoring activities of firms are lower either when the firm is a public utility or when the level of insider holdings is relatively high. This study also shows that the analysts' activities are higher for financial firms, ceteris paribus, than for nonfinancial firms, indicating that the influences of fixed-rate deposit insurance overwhelm the influences of other regulatory restrictions.

Rao and Moyer [1994] developed a theoretical model to study the role of regulatory climate in capital structure decisions of regulated electric utilities. Their model predicts that utilities will react to their regulatory climate by adjusting capital structure. They also provide cross sectional and time series empirical support for their model from their data. They do not, however, comment on the dividend policy issues of (regulated) public utilities that are an integral part of a firm's capital structure decisions.

Akhigbe, Borde, and Madura [1993] measure the common share price response to dividend increases for both insurance firms and financial institutions relative to unregulated firms. They find that insurance firms stock prices react positively to increases in dividends over a four-day interval surrounding the announcement, but that these reactions differ depending on the insurer's primary line of business. They divide the sample into these three segments: life, property and casualty, and other. Their results show that the market reaction for each segment is greater than the market reaction for financial institutions. By contrast, the market reaction for life insurers is lower than that for industrial firms, while the reactions for property and casualty firms and other insurers are both higher. However, they note that the reaction is not related to firm-specific variables like profitability, leverage, or firm size.

Finally, Collins, Saxena, and Wansley [1996] compared the dividend payout patterns of a sample of regulated firms (from banking, insurance, electric utility, and natural gas industries) with unregulated firms (from a variety of different industries). They did not find that the financial regulators' role is one of agency cost reduction for equity holders. Utilities, on the other hand, are different. They alter their dividend payout in response to changes in insider holdings. Moreover, for a given change in insider holdings. this policy change is more pronounced than the change for unregulated firms.

In summary, the literature suggests that there are different factors that determine dividend policy for regulated firms than for unregulated firms. However, not much work seems to have been done in comparing dividend policies of the two groups. In this paper some of the factors that could affect a firm's dividend policy and how they might differ between regulated and unregulated firms are studied.

The data for this study consists of 333 firms drawn randomly from editions 1-10 of ValueLine Investment Survey dated December 22, 1989 through March 16, 1990. The initial sample consisted of 500 firms, however several observations were dropped due to incomplete information on all the variables. Instead of using the data for the same years, this study uses data from approximately a decade after Rozeff's. Since his sample was drawn from the same source in 1981, an important contribution of this study is to see whether there are any significant changes in the parameter values over the time period 1981 thru 1990. The total sample is split between 235 unregulated and 98 regulated firms and that cover 56 industries. The regulated industries represented in the sample include commercial banking; savings and loan associations; investment and brokerage services; life and property and casualty insurance; electric utilities; gas; petroleum; telecommunications; railroads; and airlines industries.

The basic regression equation used in the model has the following form:

PR = B₀ + B₁G1 + B₂G2 + B₃b + B₄LSH + B₅IH + B₆IOS + e; where,

b is the firm's beta coefficient.

Numbers (information) for all of the variables are obtained directly from ValueLine Investment Survey. The database provides historical financial information for firms for several years, including dividends per share and earnings per share. It also provides a firm's growth rate over the past 5 years as well as its (Value Line's) forecast of growth over the next five years. The firm's stock ownership data, including the percentage of common stock owned by officers and directors is also included.

Past growth should be an important determinant of  the dividend payout rate. In general, a firm would look at its (recent) past growth rate when deciding how much of its earnings it needs to retain (for growth), and how much to give away as dividends. The variables G2 and IOS are proxies for the firms' growth prospects in the future. They should only be important to the unregulated firms since, regulated firms have their growth prospects determined by regulators. b proxies a firm's operating and financial risk and should be important to both regulated and unregulated firms. The number of stockholders should have a direct relationship with the payout rate for a firm. The literature provides evidence for such a relationship for unregulated firms. A priori, there is no reason to doubt this for the regulated firms as well. Lastly, variable IH proxies the level of monitoring that takes place by stockholders. This will be valuable to unregulated firms since there is no delegated monitor for them. On the other hand, because the regulator acts as a delegated monitor, the IH variable will not be as valuable for regulated firms. The variables G1, G2, b, and IOS are proxies for the transactions cost of external financing, and the variables LSH and IH are surrogates for measuring the decrease in agency cost associated with increasing the dividend payout ratio. The going hypothesis is that the variables G1, G2, b, IOS, and IH are negatively related with payout ratio [PR], and LSH is positively related with PR.

Based on the discussion in Section III A above, the following hypothesized relationship is predicted for each variable (with respect to dividend payout ratio) for regulated and unregulated firms:

For unregulated companies:

LSH should be positively related to PR; and,

For regulated companies:

The Appendix provides a correlation matrix for all the regression variables. As can be seen from the matrix, payout ratio (PR) is negatively correlated with past and future growths (G1 and G2), beta value, and insider holdings (IH). On the other hand, it is positively correlated with the number of common stockholders (LSH). However, it does not seem to be correlated with a firm's investment opportunity set (IOS).

Table 1 provides summary statistics on the regression variables for both, regulated and unregulated groups of firms. The mean dividend payout for regulated firms (around 45%) is significantly larger than that for unregulated firms (29%). Unregulated firms, on average grew more than the regulated ones in the sample over the past years, as evidenced by G1. This probably reflects the fact that regulated firms are more matured and have stabilized over time. Alternatively, it might indicate the fact that managers of regulated firms do not have as much freedom to make them grow as their counterparts in unregulated firms. The final significant difference between regulated and unregulated firms is the percentage of common stock held by the insiders. On average, the insiders of unregulated firms hold almost three times as much of their firm's common stock.

Table 2 provides the ordinary least square regression results for the whole sample (333 firms), the unregulated subgroup (235 firms), and the regulated group (98 firms). The first regression in the table demonstrates that all of the explanatory variables exhibit strong relationship with dividend payout ratio. Indeed all of them are statistically significant, most of them at the one-percent level, and the R² is 0.50.

The second regression in Table 2 reports the results for the unregulated subgroup. Indeed, all variables have the hypothesized signs, except investment opportunity set, which is statistically insignificant. The model predicts that dividend payouts of unregulated firms are inversely related to past and (expected) future growth, their systematic risk, and the level of holdings of common stock by the firm's managers, directors, and officers. On the other hand, payouts are directly related to the number of stockholders in an unregulated firm. In general, the results and their significance levels are similar to those in Rozeff [1982.

The results for the regulated subgroup provide some interesting insights regarding the payout behavior of regulated firms. First, it seems that the insiders do not play a significant role in a regulated firm's dividend policy. In the last paragraph it was mentioned that for unregulated firms this variable is negative and significant. This suggests that regulators act as substitutes for insiders in monitoring a firm's activities. If a firm is unregulated, the firm's insiders will act to "regulate" and monitor the firm's financial performance in the long run. They do not care as much about receiving dividends today if they can reinvest these dollars in positive net value projects that will increase firm value, which in turn attracts more investors.

The second difference between regulated and unregulated firms is regarding (expected) future growth (G2). This variable is statistically insignificant for regulated firms. Due to the presence of regulators, managers may feel secure in terms of capital arrangements for their forthcoming investments. Hence, they payout dividends without regard to future capital needs for growth.

Despite the fact that most coefficients of the explanatory variables reflect the hypothesized relationship (sign) with payout ratio, there are a couple of significant contrarians. For instance, for unregulated firms, investment opportunity set (IOS) does not seem to influence their payout ratio. On the contrary and counter-intuitively, this variable is positive and significant for regulated firms. Second, the future growth rate (G2) for unregulated firms has the right sign, but is only marginally significant. When studied together with the IOS variable, one could argue that an unregulated firm's payout policy is not crucially dependent upon future growth. In fact, past growth (G1) is the main "growth" that matters. Alternatively, this could be a reflection on Value Line's ability, or lack thereof, to forecast future growth correctly for most firms.

Based on the discussion in an earlier section of this paper, selected variables are dropped and regressions re-run to study the extent and the effect of the possible multicollinearity problem. Table 3 reports these results for the whole sample. As is evident from the table, there are no major changes in estimates and their significance levels when compared with the ones in Table 2. Indeed, as can be expected after reducing the number of explanatory variables, the R² value becomes smaller when we do not use the complete model.

In summary, this paper develops a model to explain dividend payout ratios of firms. Several variables employed the literature are utilized as possible determinants of dividend policy. Further, the paper investigates if any of these determinants differ between regulated and unregulated firms. For the unregulated sub-sample, results are compared with earlier studies. The usual statistical tests are carried out.

The main conclusions of the paper are that a firm's dividend policy will depend upon its past growth rate, future growth rate, systematic risk, the percentage of common stocks held by insiders, and the number of common stockholders. Moreover, the relationship is inverse in all cases except the number of common stockholders. These relationships are comparable to Rozeff's [1982] and other earlier studies on unregulated firms. More importantly, however, some of the determinants of dividend policy are different for regulated and unregulated firms. Specifically, the percentage of common stock held by insiders, and expected future growth rate, do not play a key role in a regulated firm's payout ratio.

REFERENCES

TABLE 1

Descriptive Statistics for the Regression Variables: (by Regulated and Unregulated subgroups)

TABLE 2

Ordinary least square regression results of the equation:

PR = B₀ + B₁G1 + B₂G2 + B₃b + B₄LSH + B₅IH + B₆IOS + e.

The first set of results is for all of the 333 firms (regulated and unregulated firms combined);

The second set of results is for the 235 unregulated firms; and

The third set of results is for the 98 regulated firms .

(t-values given in parenthesis)

*** Significant at the 1%-level

TABLE 3

Ordinary least square regression results of the basic equation:

PR = B₀ + B₁G1 + B₂G2 + B₃b + B₄LSH + B₅IH + B₆IOS + e.

These results are for the entire data set, however, each of the following regression estimation omits one or more variables to study the affect of the remaining variables on dividend payout ratio.

The first set of results contains all of the explanatory variables, except b;

The second set of results contains all of the explanatory variables, except LSH and IH;

The third set of results contains all of the explanatory variables, except G1 and G2; and

The fourth set of results contains all of the explanatory variables, except G1, G2, and b.

(t-values given in parenthesis)

*** Significant at the 1%-level

______________________________________________________________________________

*The author would like to thank the participants of the session of the Financial Management Association meetings where it was presented, and particularly Cary Collins and James Wansley for helpful comments. He also acknowledges a summer research grant from Mercer University's Stetson School of Business and Economics.