"3“.SNZJ‘S.‘ . 215% §';‘A u“ r ,‘ ~. efih , . “E" ,3. "9.. A .4) ~ \ spill. ' . x .n -u y .334 - .. (fisz-‘I-‘un. . 2-... :m‘: .514 . :*1 . My ”.1 "7-... 30's.”:- :JJ.- V . Jinx)”,- ' 7‘ ' "-10:31 I. : x 0 ‘5 “m v ‘v f it b‘ ‘Ptl .1} QA- ' 42:36“ ‘1“ M. I” _~¢.;.. u... ., “f. 'L 5: ~‘ltu' 5: J". .;i’—\ "‘ "0 ‘ ffé; ‘93:; 5: 3512'- : .wh "” 171; "m {EEEW Ct?“ t‘z‘a‘f «‘1th g .' '1 ‘ { Nirhrhu‘ .1“: J. 4t, 1‘7’ L 2‘ Eng.“ ”‘1‘ "A; ~ .1'1‘ 1' in“. H;H’l(—JU~7 mg 1 8' ‘x’Z L 7/ a iiiiiiigiiigiiiiwi 1312 f f l. IBMRY “hum “at. Unhoulty i This is to certify that the dissertation entitled An Empirical Investigation of Dividend Reinvestment Pians presented by Janet Marie Todd has been accepted towards fulfillment of the requirements for Ph.D. d . Finance egree 1n MXW Major professor mail/in“ “My, me a MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE 43%| Ki— —’1 t WI. WJL II AN EMPIRICAL INVESTIGATION OF DIVIDEND REINVESTMENT PLANS BY Janet Marie Todd A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Finance and Insurance 1992 III ABSTRACT AN EMPIRICAL INVESTIGATION OF DIVIDEND REINVESTHENT PLANS By Janet Marie Todd This work examines the features of a dividend reinvestment plan which attract participation and what effects dividend reinvestment plans may have on portfolio return and on the ex-dividend day price drop of the utility industry. A survey was administered to firms with dividend reinvestment plans. The data obtained from the survey show that the utility industry is the primary user of the discount and the new issue features. It also displays the increased level of participation a firm can experience when a discount feature is offered. Regression results on this data reveal that participation rates are positively related to a stock's return, current and past. Trading volume in the firm's stock is not affected by the participation level in dividend reinvestment plans. Portfolio performance in the utility industry, studied by evaluating portfolios over the years 1972-1989, is not affected by the use of a dividend reinvestment plan regardless of the discount feature. Utility portfolios cm ei' pli Th an se di th V1 he 81 h. containing dividend reinvestment plans do not outperform either the market or similar portfolios of stocks without plans. Support is found for the utility industry's ex-dividend price drop being greater than the value of the dividend. The disappearance of correlation between the drop statistic and the dividend yield, when the utility industry is tested separately, is reaffirmed in this study. Additionally, the dividend reinvestment plan was found to have no bearing on the ex-dividend day price drop. Those utilities with and without plans behave similarly. The discount feature offered by some plans was found to have no influence on the behavior of stocks that offer dividend reinvestment plans in either portfolio return or ex-day prices. The only influence the discount appears to have is the attraction of participants to the plan. I would like to dedicate this work to Matthew, my husband and to my children, Danielle, Gabrielle and Michelle. I thank them for their patience and their understanding of the time I spent away from them while completing this dissertation. iv Dal thf CO‘. to me ca GE 1 l ACKNOWLEDGMENTS I would like to thank my dissertation chairperson, Dr. Dale Damian for his ever willing help and guidance through this project. Also, gratefulness is extended to my other committee members, Dr. John Gilster and Dr. Michael Mazzeo for their insights and comments. Appreciation is given to all the Finance faculty members for sharing their knowledge and to the Ph.D. candidates for offering a wonderful support system, especially my good friends Lise Graham, Claudia Kocher and Rick Osborne. A special thanks is given to my husband, Matthew, for his help in solving my computer problems, but mostly for his loving support. TABLE OF CONTENTS LIST OF TABLES O O O O O O O O O O O O O 0 O O 0 LIST OF FIGURES . . . . . . . . . . . . . . . . CHAPTER 1 1.4 CHAPTER 2 2.1 CHAPTER 3 3.1 INTRODUCTION . . . . . . . . . . . . Theory of Dividend Policy . . . . . . Information Content of Dividends . . Tax Considerations . . . . . . . . . summary 0 O O O O O O O O O O O O O 0 LITERATURE REVIEW ON DRPS . . . . . . Background . . . . . . . . . . . . . 2.1a Types and Features of DRPs . . 2.1b Eligibility Requirements . . . Tax Consequences . . . . . . . . . . Advantages and Disadvantages of DRPs Characteristics and Trends of Firms With DRPS O O O O O O O O O O O O O O DRPs and Shareholder Value . . . . . 2.5a Theory . . . . . . . . . . . . 2.5b Empirical Studies . . . . . . . MOTIVATION AND PROPOSITIONS . . . . . Participation Rates in DRPs . . . . . 3.1a Participation Incentives . . . 3.1b Participation Rates due to Stock Return . . . . . . . . . vi ix xi 10 12 15 15 17 18 19 22 25 27 27 29 33 33 35 37 CH 3.2 Participation Rates and Volume of Trade . . . 39 3.2a Trading Volume with Respect to the Market . . . . . . . . . . . . . 40 CHAPTER 4 DATA . . . . . . . . . . . . . . . . . . . . 42 4.1 Survey Data . . . . . . . . . . . . . . . . . 42 4.1a Survey Data Relationships . . . . . . . 43 4.2 Volume Data . . . . . . . . . . . . . . . . . 46 4.3 Return Data . . . . . . . . . . . . . . . . . 48 4.4 The Data Used for Tests . . . . . . . . . . . 48 4.5 Participation Rate Data Usage . . . . . . . . 49 CHAPTER 5 METHODOLOGY AND RESULTS OF THE PROPOSITIONS 55 5.1 Methodology and Results of Proposition 1 . . 55 5.2 Methodology and Results of Proposition 2' . . 62 5.3 Methodology and Results of Proposition 3 . . 66 5.3a Individual Firm Analysis . . . . . . . 66 5.3b Pooled Sample . . . . . . . . . . . . . 78 5.4 Methodology and Results of Proposition 4 . . 81 5.5 Conclusion . . . . . . . . . . . . . . . . . 86 CHAPTER 6 RETURN OF PORTFOLIOS CONSISTING OF STOCKS WITH DRPS . . . . . . . . . . . . . . 88 6.1 Prior Work on Portfolios of Stocks With DRPS O O O O O O C O O O O O O O O O O 8 8 6.2 Utility Stock Portfolio Returns . . . . . . . 94 6.2a Data . . . . . . . . . . . . . . . . t 95 6.2b Methodology and Hypothesis . . . . . . 95 6.3 Results . . . . . . . . . . . . . . . . . . . 98 6.3a Arithmetic Mean Return Results . . . . 98 6.3b Geometric Return Results . . . . . . . 101 6.3c Individual Year Results . . . . . . . . 102 vii CHA APE 6.3d A Further Look at Arithmetic Mean Results . . . . . . . . 6.4 Conclusions . . . . . . . . . . . . CHAPTER 7 EX-DIVIDEND DAY DROP VALUES OF STOCKS USING DRPB . . . . . . . . 7.1 Background on Stock Ex-Dividend Day Behavior . . . . . . . . . . . 7.2 Data and Methodology . . . . . . . 7.3 Results . . . . . . . . . . . . . . 7.4 Conclusion . . . . . . . . . . . . CHAPTER 8 SUMMARY . . . . . . . . . . . . . . APPENDIX A DIVIDEND REINVESTMENT PLAN SURVEY Handy & Harman . . . . . . . . . LIST OF REFERENCES . . . . . . . . . . . . . . viii 102 113 114 114 118 120 128 129 131 132 Ta: Ta: Ta Te »? Table Table Table Table Table Table Table Table Table Table Table LIST OF TABLES 1 Summary Statistics of the Firm Survey Data (1971 - 1989) e e e e e e e e e e e e e e e e 2 Test for Statistical Differences Between the Utility and Nonutility Groups . . . . . . 3 Participation Rates as a Percentage of Eligible Shareholders in Discount DRPs Versus DRPs Without Discounts . . . . . . . . 4 Participation Rates as a Percentage of Outstanding Shares in Discount DRPs Versus DRPs Without Discounts . . . . . . . . . . . 5 Qualifying Effect of Utilities on Participation Rates 1982 -1985 . . . . . . . 6 The Trend of Shareholder Participation and the Effect of Return on Participation Levels as a Percent of Eligible Shareholders Utility Sub-Group . . . . . . . . . . . . . . 7 The Trend of Shareholder Participation and the Effect of Return on Participation Levels as a Percent of Outstanding Shares Utility Sub-Group . . . . . . . . . . . . . . 8 The Trend of Shareholder Participation and the Effect of Return on Participation Levels as a Percent of Outstanding Shares Nonutility Sub-Group . . . . . . . . . . . . 9 The Trend of Shareholder Participation and the Effect of Return on Participation Levels as a Percent of Eligible Shareholders Nonutility Sub-Group . . . . . . . . . . . . 10 Effects of Stock Return on Participation Rates Pooled Sample - Utilities . . . . . . . 11 Effects of Stock Return on Participation Rates Pooled Sample - Nonutilities . . . . . ix 45 47 58 60 64 68 71 74 76 80 82 Tan Tab Tab Ta: Ta‘: Ta‘: Ta‘: Ta) Ta Ta: Table Table Table Table Table Table Table Table Table Table Table Table 12 13 14 15 16 17 18 19 20 21 22 23 Effects of Participation on Trading Volume Pooled Sample - Utilities . Effects of Participation on Trading Volume Pooled Sample - Nonutilities Fetherston and Hill (1990) Table 4 - Industry Return SORTED BY AVERAGE RETURN PER STANDARD DEVIATION . . . Fetherston and Hill (1990) Table 4 - Industry Return SORTED BY COMPOUND RETURN O O O O O O O O O O O O O 0 Portfolio Returns of Utility Stocks Holding Period Returns . . . . . . Utility Portfolio Returns 1972 — 1989 Statistical Differences Between Portfolio Returns . . . . . . . . . Statistical Differences Between Portfolio Variances . . . . . . . . Ex-Dividend Day Drop Values . . . . Ex-Dividend Day Drop Values by Yield Quintiles No DRP and DRP . . Ex-Dividend Day Drop Values by Yield Quintiles DRP and DDRP . . . Correlation Coefficient Between the Drop Statistic and Dividend Yield . 84 85 9O 92 99 107 110 112 121 124 125 I127 Figure Figure Figure Figure Figure Figure Figure LIST OF FIGURES FLORIDA PROGRESS CORPORATION - Data Relationships. . OHIO EDISON COMPANY - Data Relationships. . ORANGE AND ROCKLAND UTILITIES, INC. Data Relationships. . ARITHMETIC MEAN PORTFOLIO RETURNS 1972-1989. . . . . . ARITHMETIC MEAN PORTFOLIO RETURNS 19.77-1989. e e e e e HOLDING PERIOD PORTFOLIO RETURNS - GEOMETRIC RETURN. . . ANNUAL PORTFOLIO RETURNS - 1972-1989. xi 50 52 53 103 104 105 106 I I I IFLI CHAPTER 1 INTRODUCTION Dividend policy and its relevance has long been argued in the literature. Some studies look at the amount of dividend, if any, that should be paid, while others focus on valuation effects of dividend policy. Another group of work looks at information content of dividends. Still others concentrate on particular aspects, such as stock dividends, extra dividends and dividend reinvestment plans. The main focus of this research deals with an aspect of dividend policy, namely dividend reinvestment plans (DRPs). This work will try to discern what entices investors to participate in DRPs, what effect these plans have on the firm and if ex-day dividend prices act differently for firms with dividend reinvestment plans than for those without. Before looking at the main thrust of this work, a review of dividend theory and the implications that dividend policy may have on firm value are presented. The purpose of this chapter is to set the working environment of the thesis and examine the prior work done in trying to solve what Fisher Black (1976) termed "The Dividend Puzzle". Cd] sh. in sa di fc ir ma 2 1.1 Theory of Dividend Policy Distribution of a firm's earnings to its shareholders can take place in terms of dividend payments or increases in share value (capital gains). Lintner (1956) conducted a field survey of 28 companies and reported that in discussions with managers about earnings dispersement "... dividends represent the primary and active decision variable in most situations" (p.97) versus retained earnings or savings. Current earnings were the main influence of dividends paid, with target dividend payout ratio and time for adjusting to that ratio as earnings change supplemental inputs to the decision. Once the target payout was set, management did not seem to defer from this ratio. The decision for the amount of dividend payment focused on should the current dividend level be changed, not by how much it should change. The investment opportunities were also not considered until the dividend was paid. Projects deemed profitable before the dividend payment would be reevaluated after the dividend payment if internal funding was no longer available and external financing was needed to fund these projects. Lintner's (1956) work seems to imply that dividend policy must have some relevance since a firm was concerned about maintaining dividend levels and often would pay dividends while externally raising funds to finance projects. However, there is no mention as to why a firm chooses a particular dividend payout ratio. 3 Gordon (1959) evaluated three hypotheses of what is purchased when purchasing common stock. He noted that investors buy ”(1) both the dividends and the earnings, (2) the dividends, [or] (3) the earnings. (p.99)" He set up three regression models to test each of these hypotheses. He found positive and significant coefficients for the dividend variable in all models. Gordon noted that while the dividend stream was of importance, it would not be inconsistent to have earnings streams important also. In fact, it could be that the earnings stream was sufficient to determine stock price and the dividend stream was unnecessary. Therefore, Gordon developed a more rigorous model to discern the effects of the dividend portion of earnings from the growth portion. He concluded that since cash flows received in the future are more uncertain, a larger discount rate can be applied to them. Thus, the larger the earnings stream retained, the higher the discount rate used to evaluate the future cash flows. Thus for a given current earnings, the price of the stock will be lower . (i.e. future cash flows discounted at a higher rate) the more earnings that are retained. In Gordon's model, the dividend stream is of importance because it is thought to be less uncertain (i.e. it is received at an earlier point in time than capital gains). Therefore, the higher the dividend stream, all else constant, the larger the stock value. 4 Contrary to the above findings are those of Miller and Modigliani (1961). Their theoretical results support the hypothesis that dividend policy is irrelevant.. The framework for their model is a world of perfect markets, rational behavior and perfect certainty.1 Miller and Modigliani (MM) argue that given the investment decision, as dividend payout increases more external capital must be raised to meet the desired investment level. Therefore, increasing current cash flow to the shareholders may be offset by their lower share of the future firm value. MM derive an expression which takes into consideration the positive effect of increased dividends and the negative effect of lowering shareholder's ownership. They find that in their framework the two effects exactly offset each other and conclude that dividend policy is irrelevant. MM further substantiate their claim of dividend irrelevance by analyzing three different valuation formulas. They look at the discounted cash flow approach, the investment opportunity approach and the dividend stream approach to firm valuation. In all three methods, they reach the same conclusion, dividend policy is irrelevant to firm value. Investors are just as willing to obtain their desired return through either dividends or increased terminal value. 1 For’a complete explanation of perfect markets, rational behavior and perfect certainty see Miller and Modigliani (1961) page 412. basi aP'P sto stc ma 3‘. is a! the in II III I 2? . 0‘ 5 Black and Scholes (1974) set up elaborate portfolios based on different dividend yields. They found that there appears to be no relationship between dividend yield and stock return. Thus if a firm were to increase its dividend, stock price may temporarily increase if investors thought management was signalling higher expected earnings. If it is found that there was no signal in the increased dividend, the stock price will retreat to prior levels. Therefore, they concluded that dividend policy of a firm was irrelevant in that changes in dividend policy would not affect stock price. Bar-Yosef and Kolodny (1976) use the Capital Asset Pricing Model (CAPM) framework to test for dividend policy relevance. They argue that implicit in the CAPM dividend policy is assumed to be irrelevant. Therefore, to test for relevance, they add a dividend payout variable to the market model. In one model just dividend payout is added. In a second, dummy variables for high and low payout are used. They find significance for all the dividend variables and an inverse relationship between return and payout. To discern whether the inverse relationship is due to the dividend payout or due to the correlation between high growth, risky firms having low payouts and low growth less risky firms having high payouts, Bar-Yosef and Kolodny do a two stage regression. First they determine the residual effect of dividend policy and then include the residual effect as the dividend variable in the extended market 6 model. Again they find significance of the dividend variable and thus conclude that dividend policy is relevant. The sign on the coefficient was negative thus suggesting an inverse relationship between dividend policy and return. Investors prefer their return in dividends and consequently are willing to accept lower returns per a given level of risk than if the return was obtained through capital gains. Litzenburger and Ramaswamy (1979) find different results than Bar-Yosef and Kolodny. They look at ex- dividend day price returns and find investors required a premium in the before tax expected return as compensation for higher taxes paid on dividend return. In fact, they find a strong positive relationship between before tax expected returns and dividend yield. Unlike Bar-yosef and Kolodny, these authors' findings do not imply that investors are willing to take lower returns per risk for high dividend yield. They imply the opposite. Shefrin and Statman (1984) look at the dividend decision problem from a different angle. They look at the investor's preference for dividends from a behavioral point of view. They argue that the individual's stage in his life cycle, his cash flow needs and his self control all determine the investor's preference for dividends. Shefrin and Statman argue that an investor will more readily spend cash dividends instead of reinvesting them. However, if the same return was earned in capital gains, the investor would have a difficult time selling capital to obtain the cash flo reg Pa‘i vhc in pa} re‘ ca: Th CE st re pe 7 flow (i.e. homemade dividends). There seems to be more regret when capital is dipped into than when dividend payments are spent. Given a choice of equal returning assets, an individual who is young and saving for retirement will want to invest in low or no dividend paying stock versus the high dividend paying stock in order to prevent himself from spending the return. On the other hand, a retired individual who needs a cash flow stream will want higher paying dividend stocks. These individuals may even be willing to pay a premium for certain stocks in order to obtain the desired cash flow stream, which allows them to maintain self-control and avoid regret. Therefore, dividends and capital gains are not perfect substitutes. Investors have a definite preference as to the form in which return is to be received. 1.2 Information Content of Dividends Whether investors prefer dividends over capital gains is still being discussed and debated in the literature. Perhaps the importance of dividend policy does not lie in investor preferences, but in the information the dividend policy is conveying. Many firms as found by Lintner (1956) practice a form of stable dividend policy: therefore changes in dividend policy may inform the market of management's assessment of future expected earnings. If dividends increase (decrease), this may be a signal of an expected increase (decrease) in future earnings. The information a: 8 content hypothesis is based on the belief of information asymmetry. Ross Watts (1973) tested the information content hypothesis by regressing future earnings changes on unexpected dividend changes and testing the relationship between the sign of the current, unexpected dividend changes and detrended future earnings. Both tests found weak, positive relationships. Watts concluded that while these tests support information content in dividends, the information content is very low. Watts also tested the stock price effect when dividends unexpectedly change. He found that there was a positive relationship between the two variables, and if the dividend change was a proxy for future earnings expectations, the information conveyed by the change is "trivial because the return from monopolistic possession of the information does not exceed transactions cost" (p. 211). Watts also noted that the dividend information may get lost in the noise of dividend policy. Dividends are calculated by multiplying earnings by the dividend payout and an adjustment factor that moves the dividend policy to the target ratio. Since the dividend is a relatively small number compared to earnings, much of the information carried in the dividend change could easily get lost. Battacharya (1979) developed a model where dividends were a signal of the firm's future expected earnings. The model assumed that outsiders to the firm had imperfect 9 information. Battacharya argued that if investors paid a higher price for stocks that have higher dividends, all else equal, there must be some information content in the dividend. Furthermore, the information contained in the dividend must be something that the earnings and annual reports cannot convey. If this were not true, investors would not be willing to pay a higher price for a high dividend yielding stock, all else equal, because dividends are penalized by a higher tax ratez. 'There must be a trade off between the paying of the higher tax with the obtainment of information that can not be conveyed elsewhere. This argument leads toward the support of an optimal dividend policy, not dividend irrelevance. Miller and Rock (1985) develop a model using the constraint that a firm's sources and uses of funds must be equal. The results of the model show that dividends and external financing are merely the opposite sides of the same coin. Earnings and dividend surprises convey the same information: an unexpected increase in earnings or dividends will increase shareholder wealth. However, an unexpected issue of debt or equity will signal bad news. The results of this model show support for the signalling hypothesis under asymmetric information. There is however a cost to signalling with dividends. Given an investment decision, the higher the dividend paid, 2Even under the current tax law, the investor gets the benefit of deferment with gains. 10 the more external financing that may be required. Thus there is a tradeoff between the signalling of good news by increasing the dividend and the cost of using external financing. Under the assumption of asymmetric information, it appears that dividend policy is relevant for its signalling ability, and that an optimal dividend policy may exist because of the cost trade-off discussed above. The difficulty with the models developed thus far is that none of them explain the differences in dividend policy across firms. 1.3 Tax Considerations As noted above, if dividends are taxed at a higher rate than capital gains, investors may prefer gains over taxes. Investors are concerned with their after tax return: therefore, a clientele effect may be present. Those in high tax brackets would purchase low or no dividend paying stocks and those in low tax brackets would purchase high yield .dividend stocks. Miller and Modigliani suggested this clientele effect in their 1961 paper on dividend policy. However they also noted that for valuation purposes, one clientele was as good as another. In equilibrium, no firm could gain value by changing its dividend policy. There might be a short period of time where value might change because of clientele shifts, but once equilibrium returned, 11 the value of the firm would once again be reflected in the share price. Brennan (1970) suggested that it may be detrimental to a firm to pay dividends because of tax consequences. Brennan's model showed "that so long as the market's effective tax rate T exceeds zero, the payments of dividends will be detrimental to the interests of all investors” (p. 426). T is a function of the tax rate on dividends and capital gains. Only if the tax rate on gains is equal to the tax rate on dividends will the effective rate, T, be equal to zero. Miller and Scholes (1978) showed that even with the preferential tax treatment for capital gains, investors could be indifferent to the firm's dividend policy. Their argument stems from the ability of investors to deduct interest on personal loans and that capital gains never have to be realized; thus, tax on capital gains is zero.’ An investor could invest funds, borrowed at the risk free rate, into an insurance policy. The investor borrows enough so that the deduction on the interest paid on the loan exactly offsets dividend income. The accumulation on the insurance policy is tax free, thus the dividend stream is essential transformed into a tax free stream. Under current tax law, this method no longer is feasible because interest on this type of loan is not tax deductible. 12 1.4 Sundry As presented above, dividend policy has been argued as irrelevant by Miller and Modigliani (1961) to relevant because of tax considerations by Brennan (1970) and back to being irrelevant by Miller and Scholes (1978). Signalling content of dividend policy has also been explored by Watts (1973), Battacharya (1979) and Miller and Rock (1985) among others. Watts finds that the signalling ability of dividends is trivial. However, Battacharya and Miller and Rock find evidence to support the signalling hypothesis. If dividends are signals, then there is relevance to dividend policy. This work will not look at dividend policy directly, but will focus onga program that can be offered to shareholders, the dividend reinvestment plan. The DRP is a feature of dividend policy that allows shareholders to reinvest their dividend into shares of the firm. Thus, instead of receiving a cash payment, the investor receives his return in fractional shares of stock and potential growth in value of these shares. If some investors are willing to reinvest their dividends and essentially have capital gain return, why invest in a dividend paying stock at all? A firm, as noted by Lintner (1956), spends a generous amount of managerial energy on dividend policy. Why spend time deciding how much to pay if investors are going to leave their portion of the earnings dispersement in the firm 13 anyway? This study will focus on the benefits to the firm and shareholders that get involved in these plans. The study is done in three parts. In the first section, a survey of companies that use dividend reinvestment plans was administered. From this survey, characteristics about participation rates and the potential effects DRPs have on the firm's return and trading volume are studied. The second section deals with an individual's investment decision. Does investing in a portfolio of firms that use dividend reinvestment plans give the investor a higher return per unit of risk than investing in alternative portfolios? The last section deals with the ex-dividend day price drop. Skinner and Gilster (1990), found an industry effect in the ex-dividend day price drop. If utility groups are separated from nonutilities, the utility groups (regardless of dividend yields) had ex-day drops greater than unity, while the nonutilities have ex-day drops less than unity. This study will look at the utility subgroup and see if DRPs contribute to the phenomena. Chapters 2 - 5 will deal with the survey section. Chapter 2 will review the literature on prior work looking at DRPs, while Chapter 3 will develop the motivation and propositions. Chapter 4 describes the data used, and Chapter 5 discusses the methodology and results of the propositions. Chapter 6 explores the section dealing with investment into portfolios containing firms offering DRPs. Chapter 7 14 looks at the ex-dividend day effects. Chapter 8 summarizes the three sections and reviews the major conclusions of each section. CHAPTER 2 LITERATURE REVIEW ON DRPS 2.1 Background Dividend reinvestment plans (DRPs) allow an investor the ability to reinvest his dividends in shares of the company's common stock instead of receiving cash. Corporations started these plans as a way to allow individual investors, typically small shareholders, to increase their holdings at a reasonable cost. Brokers' commission fees can often render the purchase of a small number of shares, an odd lot, to be cost prohibitive. Through a company's dividend reinvestment plan, individual investors can bypass the broker's fees, while slowly increasing their percent ownership of the firm. These plans however are not limited to small shareholders. Most firms allow all shares registered in the investor's name (individual or institutional) to participate. In 1968, First National City Bank of New York (currently Citibank) administered the first dividend reinvestment plan which was initiated by Allegheny Power. According to Davey (1976), corporations that have DRPs cite several objectives of their plans. The first and often most important to a company is the improvement of 15 16 shareholder relations. This objective is met because shareholders can increase their ownership conveniently and at low or no commission costs. Shareholders also receive the benefit of custodial and record keeping services, full utilization of dividends reinvested and/or cash contributions through fractional share ownership, and the benefits of dollar cost-averaging3 of the investment. Another goal of DRPs is to stabilize the stockholder base. Those investors who participate in DRPs are thought to be long term investors. Thus by the company attracting buy and hold investors through DRPs, the company has a stable stockholder base and potentially a less volatile stock. A third objective is the increase in market support for the company's common stock. By reinvesting dividends, there is a steady demand, at least quarterly, for the purchase of the firm's stock. Managers feel this influx of purchasing shares quarterly may help modify wide swings in the firm's stock value or at least provide a level of market value support. A fourth objective deals with plans that use new issue shares in the DRP. By using authorized but unissued shares, the firm has an influx of new capital. This new capital can be used for investment purposes and also can reduce the firm's reliance on other external sources of capital, such 3 Dollar cost averaging is a way to buy shares at a lower average cost than the average purchase price of the shares. 17 as debt. For firms with high debt levels, the influx of new capital through DRPs may be beneficial. A final objective is the potential for a reduction of administrative costs. This last objective seems to be the least prominent. For firms that use a third party (transfer agent) to administer the DRP, the reduction in cost of dividend checks and stock certificates that are mailed is often only a small savings when netted from payments for the transfer agent's services. 2-1a WW There are three basic types of DRPs. The first is the reinvestment of dividends which entails the purchase of whole and fractional common shares. The second is dividend reinvestment as in the first but with the ability to add cash contributions. Each firm sets its own minimum and maximum amount of additional cash contributions as well as how often the additional cash can be contributed. Some plans allow a monthly contribution of cash, while others . allow quarterly contributions. On average, allowable cash contributions range from $1000 to $25,000 per quarter (Wilson 1982). The third type of DRP allows investors to receive the cash dividend on all or a portion of the shares that they own, while simultaneously participating in the cash addition feature of the plan. For investors who split their share participation, two accounts are maintained, one with shares 18 that are participants in the DRP and one with shares that are not participants. Shares purchased through dividend reinvestment plans in its early years were purchased on the open market. In 1972, Long Island Lighting Company introduced a plan in which new issue shares were purchased. These plans allow firms to raise capital through their dividend reinvestment plans. In 1975, AT&T offered a 5% discount from the market price for shareholders participating in its dividend reinvestment plan. Many other firms followed ATaT's lead. The discount offered typically applied to reinvested dividends and not to cash contributions. Currently, many firms have discontinued the discount feature of their plans. 2.110 W ’ To be eligible for the DRP of most companies, a stockholder is only required to own one common share of the company. The plan is open to individual as well as institutional investors. A few firms, for example, Minnesota Power & Light, allow nonshareholders (specifically customers) to join the DRP with a minimal cash contribution. This feature provides investors with an inexpensive way to initially invest in the firm. Some firms extend the reinvestment right of dividends to preferred shareholders and recently some have extended the right of interest reinvestment to bondholders (Fredman and Nichols 1982). The reinvestment of interest payments by 19 the firm's bondholders may be attractive to corporations because the interest payment is tax deductible and also reinvested into the common stock of the company. Stocks that are held in street name are not eligible to participate in the plan. For large companies, many of their shares are held in street name causing the number of participants to be a small percentage of total outstanding shares (e.g. Eastman Kodak approximates that 80% of their stock is held in street name). 2.2 Tax Consequences Prior to the Economic Recovery Act of 1981 (ERTA), any dividends reinvested in the company were taxed as if the investor had received a cash dividend. The tax basis for stock purchased through DRPs was the market price at purchase. The discount from market price, if offered, was also taxed as income to the shareholder. Therefore, shareholders received no cash but paid taxes as if they had. In 1978, the Committee for Capital Formation through Dividend Reinvestment, a group of electric utility officials, funded a research study to be done by Robert Nathan Associates investigating the potential benefits of a tax-deferred status for new capital DRPs. Two bills supporting tax deferment of new capital DRPs, one from Representative J.J. Pickle of Texas and the other from Senator Gaylord Nelson of Wisconsin, were pending. The Nathan Report examined the potential effects of these bills 20 passing Congress. The results of the study as reported by Fredman and Nichols (1982) show the tax-deferral status to be quite beneficial. In the first year, the Treasury was expected to lose $1 billion: it would break even in the second year, gain $1.5 - $2 billion in the third year and $600 million per year thereafter. The Nathan Report also concluded that national output would increase by $2.7 billion annually: business fixed income would rise $1 billion annually; approximately 50,000 new jobs would be created annually, and an annual increase of $2.5 billion invested in new capital DRP would occur. The tax deferral status would help capital intensive industries that rely on external financing and also lower the firms cost of capital through the use of a less costly equity source than public offerings. Investors would also have an incentive to reinvest dividends instead of potentially consuming cash dividends, thus reducing inflationary pressure. Considering the Robert Nathan Report and other expert testimony, a tax-deferral status was included in ERTA, code 305(e). However, the tax law did not encompass all dividend reinvestment plans. Only investors who reinvested dividends in qualified utilities were able to benefit from the tax deferred status. 21 As reported by Boyles and Kramer (1982) and Johnson and Weber (1982) those utilities that qualified‘ needed to use authorized but unissued common stock for their reinvestment plans. The shares needed to be designated by the board as issued specifically for purposes of code 305(e). The number of shares issued to an investor in the plan was determined on a value between 95% and 105% of the common share's value during the period immediately before distribution. The Treasury department had the task of setting the time period before distribution and the formula for the share's value (e.g. the average price of all shares purchased during the dividend reinvestment period). Investors in these qualified utilities that elected to receive their dividend in stock instead of cash could defer $750 if filing separately and $1500 if married filing jointly. The dividend would not be taxed until the shares were sold, thus the deferment of taxes and taxation at the capital gains rate is obtained. The stock however must be held for at least one year after the date of distribution in order to qualify for capital gains. If the investor sold other common stock of the qualifying utility during the one year holding period, it was treated as if he had sold the qualifying stock (purchased with the reinvested dividend) and would be taxed at the ordinary rate. Unlike stock dividends, whose tax basis is based on a formula, stock ‘ See Boyles and Kramer (1982) for full details of qualification requirements. purcha Before was ti 1982 1 (1990 durin disco numbe to 15 Bake} prev. OPPO Addi DRPs beg; Pre fir the 22 purchased through the DRP had a $0 basis for tax purposes. Before ERTA, the tax basis of stock purchased through DRPs was the market value of the stock at the time of purchase. This tax preferred status only lasted from January of 1982 until December of 1985. As reported by Baker and Meeks (1990), the level of shareholder participation increased during the deferral period and remained high after the discontinuation of the deferred tax status; however, the number of firms that offered DRPs declined by 10% from 1981 to 1985. The decline continued even further after 1985. Baker and Meeks (1990) feel the other benefits of DRPs prevented investors from abandoning DRPs as investment opportunities after the tax preference was discontinued. Additionally, they believe that the decline in number of DRPs throughout the deferral period came about because firms began offering plans in 1980 in anticipation of the tax preference. When the provisions of ERTA became known, those firms that did not qualify may have begun discontinuing their plans. 2.3 Advantages and Disadvantages of DRPs Pettway and Malone (1973), Fredman and Nichols (1980, 1982), Baker and Seippel (1980, 1981) and Baker and Johnson (1988) conducted surveys of firms that had dividend reinvestment plans. Each of these authors as part of the survey asked the financial managers or chief executive officers what they felt were the advantages and 23 disadvantages of the company's DRP to the company and to the participating shareholders. Below is a compilation of the responses. es e * DRPs improve stockholders' goodwill. * DRPs create buying pressure for the firm's stock, thus there is some level of market support. * DRPs save operating expenses normally incurred with cash dividend payments. ‘ * New capital DRPs reduce dividend cash outflows. * DRPs encourage small shareholders to own larger portions of the company. * New capital DRPs improve the firm's access to capital markets by decreasing the debt ratio. * DRPs provide a regular source of equity if new issue shares are used for the plan. 2W * DRPs cause an increase in administrative problems. * New capital DRPs may cause dilution of EPS. * The DRP may become a fixture of the company even if the plan is no longer attractive to the company. 24 AdYAnLAQ2E_QI_DBEE_L2_EAI!1£12§§1B9.§h§£§h21§§2§ * DRPs provide investors a convenient way to invest small dollar amounts of dividends. * DRPs allow investors to purchase small amounts of shares at reduced service fees and brokerage fees. Often all of the fees are absorbed by the company. * Investors can increase their percent ownership of the firm. * DRPs allows participants to take advantage of dollar cost averaging. * Companies or the transfer agents do all the record keeping and custodial services for no or minimal costs. D1EQQYAnIQ92_Q£_DB£§_LQ_E§ILIQIPQDL§ * Except in 1982-1985 in qualified DRPs, participants must pay taxes on the dividend as if it was received in cash. * Participants have reduced control over the price of the shares in which their dividend is invested. * Participants need detailed records for tax purposes. * DRPs that purchase shares on the open market may limit diversification of participating shareholders. * DRP may decrease shareholder liquidity. 2.4 ' to of indus part to o avoi 18V( Sig lar low the pa: fox gr; Pa Th ir 25 2.4 Characteristics and Trends of Firms with DRPs Pettway and Malone (1973) found that a firm's decision to offer a DRP often depends on whether other firms in the industry offer the plan. If similar firms offer a particular service to their shareholders, the firm will want to offer similar benefits to its shareholders in order to avoid a potential loss of investors to the other firms. They also found that size, price/earning ratio and leverage of the firm in industries likely to have plans were significant indicators of which firms were likely to have large shareholder participation. The larger the firm, the lower its leverage and the higher its price/earnings ratio, the more likely the firm was to have high shareholder participation. In their sample, Baker and Seippel (1980) found that firms with DRPs tended to be high payout, low growth firms. Baker and Johnson (1988) found interesting trend patterns and attitude changes concerning the use of DRPs. The study looks at the differences between surveys conducted in the 19703 and the authors' most recent survey. First, managers no longer feel that firms not offering DRPs will be pressured to do so just because other firms in their industry offer them. Second, utilities, who were large users of new capital DRPs, have been using DRPs less as a source of new equity capital and nonutilities have been using them more. Utilities, however, still use new capital DRPs to a greater extent than nonutilities. Third, part1! autho marke the a inves The disc in . DRP 50m ter dix dix thl Co: Pl 26 participation rates have increased significantly. The authors suggest the reason may stem from successful marketing of the plan. The authors also noted earlier in the article that increased participation by tax exempt investors such as pension plans have caused shareholder participation rates to increase. Especially if the firm offers a discount, the purchase of the stock by the tax-free institution may be in the best interest of its beneficiaries. Lastly, the use of the discount feature has declined. The author feels the rationale for the decline in the use of discounts is difficult to explain. However, as was reported in the Wg11_§t;ggt_ggg;ngl on 6/11/90, AMAX Inc. amended its DRP to eliminate discounts in order to reduce the abuse by some large institutional traders. Institutional traders tended to buy large blocks of stock shortly before the dividend is declared, take advantage of the reinvestment of dividends at a discount and sell the stock shortly thereafter, which drives the stock price down. Many other companies, especially banks, have also been changing their plans to eliminate the discount. The downward movement of the stock poses an interesting question. Is the downward movement reflecting the ex- dividend price, or is the downward movement greater than the ex-dividend price? If the market is truly efficient, it seems plausible that the market would anticipate the action of the institutions, thus the pre-dividend price would 27 reflect the institutions actions, and the downward price movement would stem from the stock going ex-dividend. 2.5 DRPs and Shareholder value This section reviews the theory and empirical work that examines the effect of DRPs on shareholder value. The impact of a discount in the plan as well as the announcement effect of plans with and without discounts on shareholder value are studied. 2.5a Ibsen: Reilly and Nantell (1979) consider the value of the discount offered in many dividend reinvestment plans. They contend that in order for investors to be enticed to invest in the firm through DRPs, there must be some value. However, they find the value is not stemming from new wealth. It appears that there is a transfer of wealth from shareholders who do not participate to shareholders that do participate in the DRP. They derive an equation that shows if there is a discount and there are less than 100% but greater than 0% participants, the wealth change for the participants is positive. To discover the source of this wealth change, Reilly and Nantell derive the wealth change for nonparticipants. They find that with a discount and some level of participation between 0% and 100%, the wealth change is negative. Moreover, this decrease in wealth for 28 nonparticipants is equal to the wealth increase of participants. Therefore, Reilly and Nantell contend that there is no new wealth created, just a transfer of wealth. It is their belief that as the wealth change is better understood, more investors will participate in discount DRPs or shareholders will force a discontinuation of the discount plans. According to Reilly and Nantell, management that considers the beneficial equity raising ability of the plan need also realize that they are benefiting one group of shareholders over another. On the other hand, Hansen, Pinkerton and Keown (1985) develop a model that shows the discount can be treated as a flotation cost. They use Myers' (1974) adjusted present value method to determine whether discount DRPs should be used for investment. Hansen, Pinkerton and Keown argue that "[be explicitly incorporating the net discount cost into the DRP investment rule..., management is effectively impounding into the DRP project the requirement that the project's earnings be sufficient to recover the explicit value of reinvested dividends plus the implicit cost imposed upon current shares in raising the new capital. The discount is therefore correctly viewed as a flotation cost..." (p.8) Their model also shows if managers view the adoption of the DRP as an investment decision, there are no wealth transfers from nonparticipating shareholders to participating shareholders. Finnerty (1989) developed a stock valuation model and used it to study the cost of equity capital associated with new issue DRPs. If there was no discount offered for the DRP,k1 cost c hoveve DRPs 1 model cost exper a tr; Finn thei Eij (19'. Dari How par frc to in 29 DRP, he found that the cost of new issue DRP was between the cost of retained earnings and the cost of new share issuance However, if a discount was offered, the cost of new issue DRPs could exceed the cost of new share issuance.‘ In his model, Finnerty does not treat the discount as a flotation cost for the shareholders as a whole, but treats it as an expense of the nonparticipants. Finnerty also develops a model to determine if there is a transfer of wealth from nonparticipants to participants. Finnerty compares the nonparticipants' decision to retain their dividends to the allowance of letting their rights expire in a rights offering. Along with Reilly and Nantell (1979), Finnerty finds that if the discount is zero or participation is 100%, there is no transfer of wealth. However, if the discount is greater than zero and participation is between 0% and 100%, a transfer of wealth from nonparticipants to participants occurs. 2.5b Wigs Peterson, Peterson and Moore (1987) did an event study to determine whether the initial filing of shares to be used in new capital DRPs caused a firm's stock to have abnormal returns. Since previous studies, such as Asquith and Mullins (1986), Masulis and Korwor (1986) and Mikkelson and Parch (1986), have found negative abnormal returns around the issuance of new equity and equivalents, Peterson, Peterson and Moore (1987) also expect to find negative 30 abnormal returns. They note however that the tax deferment of qualified utilities during May 1981 - December 1985 may offset or more than offset the negative effect on equity value that comes from new issuance of stock. To test their hypothesis, they split the sample into nonutilities, utilities that file before May 1981 (no tax benefits) and utilities that file after July 1981 (with tax preference)”. The results indicate that for nonutilities, there is no significant abnormal market reaction around the date of. filing. The utilities adopting the new issue plan before the tax benefit period experience a negative abnormal return the day after the filing. Utilities that filed during the tax preferential period had a nonsignificant negative market reaction. Because of the uncertainty of the public date of the filing, a zero to ten day cumulative abnormal return was calculated. The utilities in the period of qualification have a significantly (5% level) different market reaction than the nonutilities and pre-May 1981 utility group. In fact the abnormal cumulative return is positive suggesting benefits from the tax deferment outweighing the negative signalling of new issues. If new issue shares are used for DRPs, leverage will be lowered and according to Masulis (1980), on average, a 5 Even though the actual deferral period does not begin until 1982, July 1981 is used for the separation date between tax deferred and non-tax deferred time periods because the law was passed at this time, thus the market should already be reflecting the expectations of the tax law. decreas mnmfs fact in Myers a (1984) inform stock) tinanc finan: both , are e Purch Dubot manag they the ; that the wet} div; The the Con ahr mo\ pCH 31 decrease in leverage will adversely affect stockholders. Dubofsky and Bierman (1988) argue that new issue DRPs may in fact increase shareholder value. Their argument stems from Myers and Majluf (1984) and Myers (1984). Myers and Majluf (1984) show that external equity financing signals negative information about the firm (overvaluation of the firms stock), and Myers (1984) develops a pecking order of financing where internal financing is a better choice of financing than external. New DRPs have characteristics of both internal and external financing methods. New shares are external financing but having current shareholders purchase these shares is a form of internal financing. Dubofsky and Bierman's (1988) argument is that perhaps managers are signalling that shares are undervalued, thus they offer them to current shareholders in order to benefit the shareholders. If this argument is true, then it follows that there should be an increase in value associated with the announcement of new issue DRPs. Dubofsky and'Bierman used a market-adjusted returns . method to determine whether the announcement of discount dividend reinvestment plans were responded to favorably. Their findings conclude that the market reacts favorably to the announcement of discount DRP. Dubofsky and Bierman conclude the positive abnormal returns around the announcement date indicate that 1) the sampled firms are moving to a more optimal capital structure, 2) there is positive information conveyed about the undervaluation of 32 the firms' stock price and/or 3) the costs associated with the plan are lower than the transaction and brokerage costs associated with purchasing shares or issuing new shares. A problem with Dubofsky and Bierman's test is that while they use only firms with discount features in their DRPs, they do not distinguish between those firms using new issue shares or market shares for their plan. Their conclusions seemingly are based on the assumption of new issue plans with discounts. The favorable response seems to be stemming from the ability of shareholders to purchase the shares at lower commission costs and at a discount. As can be seen by the above literature review, the effect of DRPs (discount and/or new issue) on shareholder value is inconclusive. Also, the work done on DRPs thus far has not addressed the potential long term effects that these plans may have on the company. terms effec retui long 3.1 inst divj Casi uti: inc. CHAPTER 3 MOTIVATION AND PROPOSITIONS This section develops the motivation of the study in terms of what causes investor participation in DRPs and the effects of participation on the firms' stock in terms of return and volume of trade. The focus of the study is on long term trends. 3.1 Participation Rates in DRPs Although participants in DRPs essentially receive stock instead of cash for their dividend, the amount of the dividend is subject to taxation as if it were received in cash. Except for participants in DRPs of qualified utilities during 1982-1985, there appears to be no tax incentive to participate. According to Davey (1976), DRPs are often used to raise new capital. Baker and Seippel (1980) find that firms offering DRPs tend to have high payout and low growth. High payout, low growth firms typically are in a mature stage and probably do not need the constant influx of equity capital that new issue DRPs provide. They would be more likely to use debt to the fullest extent thus utilizing the tax 33 benefit betvee new ca want a payout result 4, Tab | new ca intere this I in hig Capit; firms Chara iHVQs Parti inves Were the i inVes Pay 1 for 1 “9W3 34 benefits. Therefore, there seems to be a discrepancy between the findings of the two studies. Upon investigation, it appears that firms using the new capital option are mainly utilities that often need or want an influx of equity capital, but yet are typically high payout, low growth firms. Utility firms seem to satisfy the results of both studies. As will be demonstrated in Chapter 4, Tables 1 and 2 support the notion of utilities using the new capital option significantly more than the nonutilities. However, shareholders that participate in DRPs are more interested in growth return than dividend return. Under this premise, why would these shareholders initially invest in high payout, low growth firms? Investors interested in capital gains would invest in low dividend, high growth firms, all else equal. Consequently, there must be characteristics of DRPs that entice these investors to invest in high payout, low growth firms and subsequently participate in the DRP. Also, some dividend seeking investors who initially invested in these high payout, low growth firms potentially were enticed into participating in the DRP. Some aspect of the DRP must be attractive enough to cause this group of investors to forgo the dividend payment (even though they pay taxes as if the dividend was received in cash) in return for purchasing new shares. According to Baker and Meeks (1990), there has been an upward trend in participation rates in DRPs. With a lack of tax 1 types DRPs parti are t commi durir 3.1a feat‘ obta reSp the disc 22.8 Part Vhil that Way Inve take mark inst 35 tax incentives and investors seemingly investing in contrary types of firms, there must be some motivating feature(s) of DRPs or the firms that cause shareholders to want to participate. The most likely motivators for participation are the discounts offered by some of these plans, the low commission cost, the tax benefit of qualified utilities during 1982 - 1985 and the past stock returns of the firm. 3.1a WW Baker and Johnson (1988) tested the effect discount features have on participation rates using survey data obtained for the year 1986. Their test computed the mean response of the surveyed firms. They concluded that since the mean participation rate in 1986 for firms offering discounts versus firms not offering discounts was 24.9% and 22.8% respectively, the discount feature does not stimulate participation. Later, Baker and Meeks (1990) state that while participation rates are increasing it does not appear that "plan features such as discounts ... seem to influence significantly participation rates" (p.9). However, a discount feature in a DRP gives investors a way to purchase shares of the firm below the market price. Investors interested in a company would potentially want to take advantage of purchasing additional shares at below market price. Also as noted earlier in Chapter 2, institutions join DRPs to take advantage of the discount, 36 sometimes for arbitrage reasons and sometimes to benefit their constituents. Following the above arguments, it appears the discount should provide incentive for shareholder participation. Thus the expectations of this study are contrary to the above studies. This work will test the incentive that discounts may have on participation over the long run and expects to find higher participation rates in firms with discount features in their DRPs. Also, between 1982 and 1985, participants in DRPs of qualified utilities had a tax deferral benefit. Thus, participating in DRPs of qualified utilities would allow I investors to defer the tax on reinvested dividends and essentially convert the dividend return into a capital gain. This tax benefit could draw new investors to the firm who want to take advantage of the tax deferral. It also might entice current shareholders who previously were nonparticipants to join the DRP. Based on the above examination of potential motives for shareholder participation, the following propositions are put forth. Bream Firms that offer the discount feature in their DRPs will have, on average, higher participation levels than firms that do not offer the discount feature. and 37 W During the period of 1982 - 1985, utilities that qualified under ERTA, thus allowing their participants to benefit from tax deferrals, had a larger percent participation, on average, than utilities that did not qualify under ERTA. Propositions 1 and 2 have been studied before, mostly in terms of survey responses from CEOs as to what they feel are shareholder benefits or in terms of general trend patterns or average responses. This study will formally test the propositions based on actual participation rates in DRPs. The specific tests are discussed in the methodology chapter. 3.110 W Pettway and Malone (1973) regressed participation rates in DRPs of 33 industrial firms on several firm characteristics for the year 1972. The variables used were return variables, including a weighted average return for 1970 to 1973, P/E ratio and growth rate in EPS for 1963-1972 (arithmetic mean), the number of shares outstanding in 1972, current ratio, D/E ratio and average payout ratio. The only variables that were found to be significant were outstanding shares, D/E and P/E ratios. The return variables (other than P/E) were not found to be significant for the year 1972. The results for P/E ratio which were significant suggested high P/E ratios induce high participation rates. A high P/E ratio however, may not be a good in signifI opport both. potent partii of re their Halo: Sign. may futu usag thu: The: IEt For . W11 01- th( re 38 good measure of return. A high P/E measure usually signifies that investors think the firm has good growth opportunities, has a relatively safe earnings stream or both. Thus the P/E variable probably does not address the potential effect that stock return may have on participation. Hence, this study will reexamine the effect of return on participation. Consider now the investors who are interested in having their dividends reinvested. According to Pettway and Malone's (1973) result, average return variables do not significantly affect participation. However, it would seem a current high stock return or a previous high stock return may entice investors to join the DRP in anticipation that future returns will be just as good. The average return usage by Pettway and Malone may have smoothed earnings and thus lost the effect of a period of high or low return. Therefore this study proposes to look at the effect of return on participation rates over a span of twelve years. For each year actual participation rates and stock returns . will be used. It is expected that a period of high return or high expected return will encourage nonparticipants in the DRP to enroll, thus increasing the participation rate. Based on the above argument, the following proposal is made. Proposition; The participation rate in the firm's DRP is directly related to returns on the firm's stock. It importa who exp to join for the 8: comes ‘ announ conter lag e retur Patti stud; 3.2 Pla ter She St. th “1' 39 It is expected that a lagged return variable may be an important indicator of participation rates since investors who experience a high or increased return may decide ex post to join the plan. This argument assumes expectations are for the high or increasing return to continue. Since information about the firms' potential returns comes more often than annually (e.g. quarterly earnings announcements, investment opportunities, etc.), the contemporaneous return may capture some of the hypothesized lag effect. Thus it can be argued that contemporaneous return on a yearly basis may also be a good predictor of participation rates. Therefore, the return variable will be studied contemporaneously and lagged one time period. 3.2 Participation Rates and VOlume of Trade Shareholders who participate in dividend reinvestment plans would seem to be buy and hold investors, i.e. long term investors. If a company has a large percentage of shareholder participation in its plan, the firm may have a stable stockholder base. With a stable stockholder base, the volume of trading in the firm's stock may decrease. ' With decreased volume of trade, the stock's liquidity may also decrease. According to Amihud and Mendelson (1986) as liquidity of the asset decreases, the asset's return, in this case the common stock, is increased. Therefore, determining if shareholder participation is large enough to decreas« share's 3.2a 1 I! use a : relati mirror and he Velum: Other annou effec tfade mode: foll. bu 40 decrease the volume of trade which may have an effect on the share's liquidity and thus return is a question of interest. 3.2:: MW In their study, Richardson, Sefcik, and Thompson (1986) use a simple model for the firm's volume and its relationship with the market volume. The model vit = 01 + Bivmt + V1A5A + V1151 + en. (1) mirrors the market model but uses volume instead of return and has an addition of two variables. Vu = the weekly volume of firm i; Vm,= the weekly volume of market, and the other terms represent "abnormal" reactions to dividend . announcements and trading intervals, respectively. The effect that participation rates in DRPs has on volume of trade will be tested following this model. The general model used in this work for volume relationships is as follows: Vie = 0‘1 1’ Bivmt 1' 51%?“ + Cu. (2) where V“ = the annual trading volume of firm i in year t. ‘Wm a the annual trading volume of the market in year t, and i %Pit = the percent participation rate in firm i's DRP at time t. DRPs may have an effect on volume due to the potential buy-hold investors that participate in DRPs. Once 41 individual investors join the plan, it appears reasonable to assume they are interested in having a long term relationship with the company and thus become a buy-hold investor. If there are enough buy-hold investors, the volume of trade in the firm's stock should decrease. Following this argument, the subsequent proposition is asserted. 2199931112114 The percent participation in a firm's dividend reinvestment plan with respect to a) the number of shareholders and b) the number of total shares will have a negative effect on the volume of trade in the firm's stock. If participants are not buy-hold but institutional investors interested in moving in and out of stocks to take advantage of discounts, the proposed volume effect may be offset or in extreme be in the opposite direction than predicted. CHAPTER 4 DATA 4.1 Survey Data Once a year, the Stand;rd_and_gggz;§_gutlggk publishes a list of companies that have dividend reinvestment plans. To identify companies that had plans for at least 15 years, the lists published on November 11, 1974 and December 27, 1989 were used. 131 firms that had DRPs in both time periods were identified. A survey was formulated (see Appendix 1) and mailed to the shareholder relations department of each firm identified. The addresses were obtained from the companies' annual reports. In some cases, the shareholder relation department was not available and the survey was sent to the Secretary or Treasurer's office. Each respondent was asked to provide the following information for each year the firm offered the plan: 1. the percent of eligible shareholders that participated, 2. the percent of total outstanding shares that participated, 3. the percent discount offered, 4. whether new issues or market shares were used for the plan and 42 43 5. whether additional cash contributions were allowed. Also asked was the month and year the plan began as well as the name and title of the correspondent. The rate of response from these surveys was approximately 25% - 30%. While the rate of response was fair, there were two problems that arose. The most obvious was the small number of firms in the sample that would stem from a 30% response rate. The other was the fact that not all of the correspondents completely filled out the survey. Many stated that the data requested was unavailable or unobtainable especially in the early years of the plan. For this reason, many of the returned surveys only reported the last three to five years of the requested data. This length of time seems too short to get any meaningful results. Thus the "useable" response rate was lower than 30%. To increase the sample size, the SLQBQQIQ_ADQ_EQQIL§ Outlook September 24, 1979 issue was used and the names of 164 more companies that had plans in 1979 and 1989, a ten . year window, were obtained. A survey was sent to each of these companies and approximately the same response rate was obtained, again with some of the returned surveys having an insufficient amount of data reported. 4.1:: W The "useable" sample consists of 56 companies, 33 utilities and 23 nonutilities of which two are financial 44 institutions. The characteristics of the sample are analyzed and presented in Table 1. Let 1 if firm i has characteristic 2 x u: = 1 0 if firm i does not have characteristic 2 1 if firm i has characteristic 2 xisz = 0 if firm i does not have characteristic 2 where: u = utilities 3 = nonutilities z = discount offered: new issue shares used: cash contributions allowed It appears from Table 1 that the utility group utilizes the discount and new issue features more readily than the nonutility group which may cause higher percent participation in utility DRPs. According to the central limit theorem, a sample size of about 30 is enough to ensure that the distribution can be approximated by a normal distribution; however, for binomial distributions, the normal distribution approximation is quite good for sample sizes of about 25. Therefore, to test the hypothesis Ho: u... = u... versus H.= u... v‘ flu where pu,= the mean value for the utility group offering characteristic 2, and 45 Table 1 Summary Statistics of the Firm Survey Data (1971 - 1989) ‘ Characteristic Utility Nonutility Number of firms 33 23 Proportion of firms 0.576 0.261 offering a discount* Proportion of firms that 0.879 0.478 may use new issue shares* Proportion of firms that 1.00 1.00 allow cash contributions* * Many of the firms offered the characteristic only during a portion of the sample period. 46 u“,= the mean value for the nonutility group offering characteristic 2, a normal distribution was assumed and t-tests were performed for each characteristic. For each group, firm i will offer 2 with probability p. Therefore E(p) = p. Since p is unknown it will be estimated by Y/n, where Y equals the number of firms in the group that offer 2, and n equals the total number of firms in the group. Table 2 shows that firms in the utility group offered a discount and used new issues in their dividend reinvestment plans significantly more than the firms in the nonutility group. In both cases, the p-value was less than 0.005. The results of Table 2 support the separation of the sample into two groups for the testing of propositions 1 -40 4.2 Vblume Data Yearly volume of trade data for the market was obtained from the W and The MW Yearly volume for the year of 1989 was obtained from the Wall Street Journal January 2, 1990, page C2. All of the firms in the sample are listed on the NYSE; however, most of the firms are also traded on the AMEX. Therefore, a sum of the NYSE and AMEX volume of trade in each year is used. A few of the survey responses provided data beginning in 1971; 47 Table 2 Test for Statisical Differences Between the Utility and Nonutility Groups HO: ”“2 = ”82 Ha: Pu: " I‘sz Qisgoun; Featuue LLtilitx Won t' ' p z p = Y/n = 0.576 p z p = Y/n = 0.250 02 z pq/n = 0.00740 02 z pq/n = .00839 Test Statistic: t = pu - pa - 0 [(pq/n). + (15>CI/r1>..]”2 t = 2.5068 df = 54 p-value < .005 E§E_I§§B§_H§§Q§ 93.11131 Neon—1:11.151 pzz p = Y/n = 0.879 pzz p = Y/n = 0.478 a z pq/n = 0.00322 0 z pq/n = .01085 Test Statistic: t = Pa"1% ' 0 [(pq/n). + (pq/nm‘” t = 3.3806 df = 54 p-value < .005 48 thus, volume data for the NYSE and AMEX is obtained from 1971 to 1989. Quarterly volume of trade data for the individual firms is obtained through the Qgu_gungfi_flgu§_3guzlgy§l_figzulg§. The earliest date for which the data is available is the first quarter of 1978. The volume series is the composite volume of trade for each of the firms. The data series, which runs from 1978 until 1989, is then annualized by taking the sum of the quarters in each year. 4.3 Return Data The 1990 QBSB tape is used to obtain return data for. each individual firm. The length of the obtainable volume series seems to be the limiting factor for data length; therefore, monthly data from 1978 to 1989 are obtained. The monthly returns are annualized as follows. Let kt“ = the monthly return for firm i in year t. Let k“ - the yearly return for firm i in year t. Then, 12 kit a ‘l' (1 + kijt) " 1° (3) i=1 4.4 The Data Used for Tests The results of Table 2, the regulation of utilities and the tax benefits offered during 1982 - 1985 for qualifying utilities suggest that the utility groups will behave differently than the nonutility group. Therefore, the sample of 56 companies, 59% of which are utilities and 41% 49 nonutilities, will be tested as two separate groups. The length of the data set used for the empirical studies runs from 1978 to 1989. However, some of the firms only have survey data for ten years instead of the full twelve. A missing data variable is used for the survey data that was unobtainable. Ralston Purina Company did not supply information on whether they offered a discount in their plans. Therefore, for any tests that include discount as a variable, Ralston Purina company was omitted. Kerr-McGee did not supply participation rates in terms of outstanding shares or if they offered the discount option: therefore, in tests using these variables, Kerr-McGee was omitted. Sun Company did not supply information on participation rates in terms of outstanding shares: therefore, Sun Company is omitted in tests that use this variable. 4.5 Participation Rate Data Usage When participation rates in DRPs are used in a model to test a proposition, the model was run twice, once with participation rates in terms of eligible shareholders and once with participation rates in terms of outstanding shares. The justification for running the model twice can be seen in Figure 1. When the firm first offers the discount both measures of participation rates increase, however when the firm discontinues the discount, the participation rate in terms of outstanding shares drops more 50 0.5 0.45 '- 0.4 - 0.35 '- 0.3 '- 0.25 P- 0.2 "' 0.15 '- 0.1 r- 0.05 '- 2 3 4 3 A 0 1 L/l l l l L\A 5 L 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 Year D I manholder pert. + I share out 0 5 discount Figure 1: FLORIDA PROGRESS CORPORATION - Data Relationships. This figure presents percent participation rates in the DRP in terms of eligible shareholders and outstanding shares. It also shows the percent discount offered. 51 than the participation in terms of eligible shareholders. The larger drop in outstanding share participation is also displayed in Figures 2 and 3. This phenomena is potentially the result of the type of investors participating in the plan. If a company discontinues the discount in its DRP, the investor may drop out of the reinvestment plan because there is no longer any benefit (e.g. the purchase of shares at a discount) in which to take advantage. Since institutional investors typically have larger holdings in companies than individuals, it would be expected that the participation rates in terms of outstanding shares would drop more than participation rates in terms of eligible shareholders if institutions as well as individuals dropped the plan. The reasoning is that individuals as well as institutions each count as one eligible shareholder regardless of the number of shares they own, but differ substantially in their weight in terms of percent of outstanding shares. Therefore, if individuals drop the plan, a small change in both eligible and outstanding participation rates occurs. However, if an institution drops the plan, a small change in participation rates in terms of eligible shareholders occurs, but a potentially large change in the participation rates in terms of outstanding shares occurs. 52 0.3 '- 0.25 " 005- O 0 1 L 1 l 1 1 [\A L l L i 1978 1979 1980 1981 1Q! 1903 1Q4 1985 1986 1!? 1988 1989 <> 4» o 4» 0 o Your 0 I shareholder pm. + I then out 0 I discount Figure 2: OHIO EDISON COMPANY - Data Relationships. This figure presents percent participation rates in the DRP in terms of eligible shareholders and outstanding shares. It also shows the percent discount offered. 53 0.32 0.3 - "' 0.28 - - 0.26 '- " 0.24 '- 0.22 " 0.2 - 0.18 - t 0.16 '- E 0.14 - 1' 0.12 '- 0.1 '- 0.08 *- 0.06 '- 0.04- - 0.02 -' L 0 0 0 0 > > l l l l l l A L 1 L J. v 1978 1979 1980 1981 1982 1983 1984» 1985 1986 18,87 1&8 1989 Your D x shaoholdor port. + a: sham out 0 X diocwnt Figure 3: ORANGE AND ROCKLAND UTILITIES, INC. - Data Relationships. This figure presents percent participation rates in the DRP in terms of eligible shareholders and outstanding shares. It also shows the percent discount offered. 54 To allow for the possibility of the different types of investors to have a distinct effect on the variables being tested, each model using participation rates will be run twice as discussed above. CHAPTER 5 METHODOLOGY AND RESULES OF THE PROPOSITIONS This chapter will discuss the testing of the four propositions. The specific methodologies used and results of each individual test are presented as well as a combined summary of all the proposition results. 5.1 Methodology and Results of Proposition l W Firms that offer the discount feature in their DRPs will have, on average, higher participation levels than firms that do not offer the discount feature. To test this proposition, the average participation rates over the time period were calculated for all firms. The firms were then ranked in ascending order, with 1 the rank for the lowest average participation rate. The firms were then divided into two groups, firms that offer a discount, sample A, and firms that do not, sample 8. Since the underlying distribution is unknown, a Mann-Whitney U test was done to determine whether the group that offers a discount has a higher average rate of participation than the 55 56 group that does not. The same test is conducted within the utility and nonutility groups. The U statistic was calculated as UA = n11'12 + ["1011 + 1)]/2 " WA (4) where In = the number of firms offering a discount. I» = the number of firms that do not offer a discount. w, -- the rank sum for sample A. For sample sizes greater than or equal to ten, a normal approximation can be made and the following z test statistic calculated. Z = [UA - E(UA)]/au. (5) where E(UA) = [n1n21/2 (6) V(UA) " [111112011 '1' n2 '1' 1)]/12 (7) The test is used to determine whether the distribution of sample A lies to the right of sample B. Stated formally, Hypothesis 5-1 Ho: There is no difference in the distribution of participation rates in firms' DRPs that offer discounts and those that do not. 57 H,: The average participation rate in firms' DRPs offering a discount is greater that firms that do not. (A's distribution lies to the right of 3's distribution.) Since the rank sum for A, W,, is expected to be large, U, is expected to be small. Therefore, the z-statistic should be less than or equal to -1.64, the Z-value for a one sided test at 5% significance level. For the nonutility group, the number of firms that offered a discount was only seven; therefore, for this subgroup, the U statistic was compared to a U;'value obtained from a table of U-values for a 5% significance level. If UA is less than Uo, then the null hypothesis can be rejected. Tables 3 and 4 show the results of the Mann-Whitney U test. Table 3 looks at the participation levels as a percent of eligible shareholders. Table 4 looks at the participation levels as a percent of outstanding shares. As . expected the ug‘value was small in the three subgroups and the z-statistic significant at the 5% level for all firms and the utility subgroup. For the nonutility subgroup, the (h value was smaller than the specified Og‘value. The results of these tests support Proposition 1. For firms with DRPs, on average, participation rates as a percentage of eligible shareholders or outstanding shares is 58 Table 3 . Participation Rates as a Percentage of Eligible Shareholders in Discount DRPs Versus DRPs Without Discounts A: Firms in the Sample that Offer a Discount' B: Firms in the Sample that Do Not Offer a Discount H H o: No difference in the distributions ll: A's distribution is shifted to the right of B's (Firms that offer discounts have higher participation levels) under H., UA:st% or Z s -1.64 (at 5% significance) where: UA = “Na '1’ ["1131 '1' 1)]/2 " WA Z = [011' E(UA)]/UUA All Firms: A B Discount Offered No Discount n1 = 26 n2 = 28 w, = 879 wB = 606 UA = 200 UB = 728 E(UA) = n1n7/2 = 364 a“ = {[n1n2(n1+n2+1)]/12}1’2 = 57.76 2 = - 2.839‘ Utility Subgroup: A B Discount Offered No Discount n1 = 19 n; = 13 w, = 355 wB = 163 In = 72 Ln = 175 E(UA) = n1n2/2 = 123.5 a... = {[n1n2(n1+nz+1)1/12}“2 = 26.06 2 = - 1.976‘ 59 Table 3 (cont 'd) : Firms in the Sample that Offer a Discount : Firms in the Sample that Do Not Offer a Discount 0: No difference in the distributions .: A's distribution is shifted to the right of B's (Firms that offer discounts have higher participation levels) Under 11., UA 5 U0 or Z s -1.64 (at 5% significance) where: UA = nlnz '1' [”1011 '*' 1)]/2 ' WA Z = [0A. E(UA) J/OUA Nonutility Subgroup: A B Discount Offered No Discount n1 = 7 n; = 15 w, = 107 wB = 146 UA = 26 UB = 79 U0 = 28 U, = 26 is less than Uo = 28 ' * Significant at the 5% level 60 Table 4 Participation Rates as a Percentage of Outstanding Shares in Discount DRPs versus DRPs Without Discounts : Firms in the Sample that Offer a Discount : Firms in the Sample that Do Not Offer a Discount : No difference in the distributions ,: A's distribution is shifted to the right of B's (Firms that offer discounts have higher participation levels) Under 11., U, 5 U0 or z s -1.64 (at 5% significance) where: UA = n1152 '1’ [“1011 ‘1' 1)]/2 ’ WA Z = [UA’ E(UA) 1/0UA All Firms: A B Discount Offered No Discount “1 = 26 112 = 26 w, = 955 w, = 423 U, = 72 UB = 604 E(U,) = n1n2/2 = 338 on, = {[n1n2(n1+n2+1)]/12)1’2 = 54.64 2 = - 4.868' Utility Subgroup: A B Discount Offered No Discount n1 = 19 112 = 12 w, = 370 wB = 126 U, = 48 U, = 180 E(U,) = n1n2/2 = 114 a", = {[n1n2(n1+n2+l)]/12}1’2 = 24.66 2 = - 2.6766‘ 61 Table 4 (cont'd) A: Firms in the Sample that Offer a Discount B: Firms in the Sample that Do Not Offer a Discount 1%: No difference in the distributions IL: A's distribution is shifted to the right of B's (Firms that offer discounts have higher participation levels) Under H,, tubs U, or Z s -1.64 (at 5% significance) where: UA = “1’12 '1' [n1(n1 + 1) J/Z ’ WA Z = [Ur E(UA)]/0UA Nonutility Subgroup: A B Discount Offered No Discount 111 = 7 112 = 14 w, = 125 w, = 106 U, = 1 U, = 97 U0 = 26 i U. = 1 is less than Uo = 26 * Significance at the 5% level 62 significantly greater than for firms whose DRPs do not offer discounts. These results support the notion that the discount offered does entice shareholders to participate in a firms DRP contrary to Baker and Meeks (1990) assertion that the discount feature does not significantly influence investor participation. 5.2 Methodology and Results of Proposition 2 mm: Utilities that qualified under ERTA, thus allowing their participants to benefit from tax deferrals, had a larger percent participation, on average, than utilities that did not qualify under ERTA. The utility sample of firms will be split into two groups, those that qualified for the tax deferment and those that did not. The years tested will be 1982 - 1985, the years in which the tax deferment was in effect. Of the 31 utilities, only four of them did not qualify for the tax deferment. The average participation rate during the 4 year time period is calculated for each firm and ranked in ascending order. The Wilcoxon Ranked Sum test is used to determine whether the participation rate distribution of qualified utilities is the same as unqualified utilities. The rank sum, W, and W,, are calculated for each sample by summing the ranks of each group. 63 Let. W“ = the rank sum for nonqualified utilities W5 = the rank sum for qualified utilities In a the number of nonqualified utilities, and 1% = the number of qualified utilities. The test is used to determine whether the distribution of sample A lies to the left of sample B, therefore the hypothesis is Hypothesis 5-2 H,: There is no difference in the distribution of average participation rates in qualified and nonqualified utilities. H,: The average participation rate in qualified utilities is greater than in nonqualifying utilities. (A's distribution lies to the left of B's distribution.) Since the rank sum for A, W,, is expected to be small, W, is compared to W, obtained from a Wilcoxson Ranked Sum table“. inbmust be smaller than the lower limit or larger than the upper limit in order for the null hypothesis to be rejected. The significance level used is 5%. Table 5 shows the results of the test. As can be seen in Table 5, if one used participation in terms of percentage of eligible shareholders, the null 6The table was obtained from Selected Tables in 64 Table 5 Qualifying Effect of Utilities on Participation Rates 1982 - 1985 Participation Rates Percent of eligible Percent of outstanding Shareholders Shares n1=4 w,=24 n1=3 w,=30 112 = 27 W, = 472 n2 = 27 W, = 435 W,: 35,93 9 0.0469 W,: 21,72 9 0.0429 36,92 9 0.0535 22,71 @ 0.0502 w,w, significant - reject H, cannot reject H, 65 hypothesis can be rejected. However, when using participation as a percent of outstanding shares, the test cannot reject the null hypothesis. The results are mixed. The effect of the tax deferment of reinvested dividends may show up in the percentage of eligible shareholders not outstanding shares because the allowable deferment, $750 for individual and $1500 for joint filers, benefits individual investors and not the large institutions. Large shareholders, corporations or institutional investors, already have an 85%7‘exclusion or a non-taxable position if they are pension funds, thus the deferment is not as beneficial. If individuals caused the increase in participation rates during this time frame, then it would be expected that the result will show up more readily as a percent of eligible shareholders (where the weight of individuals and institutions are equal) than in terms of outstanding shares (where large shareholders overshadow individuals). 7At the time of the law, 85% was the exclusion rate. Currently the rate is lower. 66 5.3 lethodology and Results of Proposition 3 2122951321224 The participation rate in the firm's DRP is directly related to returns on the firm's stock. 5-33 WM To test this proposition, a regression on the following equation was run on each firm individually. %Pit = 01 + BIYEARt + 61RETURN1t '1' TiRETURNit-1 + (it. (8) where %P“ a the percent participation in the dividend reinvestment plan for firm i at time t, YEAR, a the trend variable, year, RETURN“ = the contemporaneous annual stock return of firm i at time t and, RETURN“q_= one year lag in annual return of firm i. Stated formally, the hypothesis tested with the above equation is Hypothesis 5-3 H,: The firm's stock return has no effect on the participation rate in the DRP. H,: The participation rate in the firm's DRP is directly related to returns on the firm's stock. 67 The trend variable was included in this regression because it was noted by Baker and Meeks (1990) that there seemed to be an upward trend in participation rates. Therefore, the portion of participation that stems from the trend is captured by the variable YEAR and distinguished from the portion of participation that is hypothesized to stem from return. The regression was run individually for each firm in the utility group and the nonutility group. Tables 6 and 7 show the results for utilities using eligible shareholder participation rates and outstanding shares participation rates for %P respectively. Tables 8 and 9 show the results for nonutilities using outstanding shares participation rates and eligible shareholder participation rates for %P respectively. The results indicate a notable upward trend in participation rates. Table 6 shows the coefficient for the variable year was positive 96.9% of the time and positive and significant (at the 5% level) 78.1% of the time, suggesting that for most firms participation increases with time. Table 7 is not as convincing in the time trend significance. Only 48.4% of the coefficients are positive, while 29% are positive and significant. Table 8 has 76.2% of the coefficients positive and 47.6% positive and significant at the 5% level. Table 9 shows that 91.3% of the coefficients on the trend variable 68 Table 6 The Trend of Shareholder Participation and the Effect of Return on Participation Levels as a Percent of Eligible Shareholders Utility Sub-Group %P1t - a, + 6mm + 61RETURNR + riRETURNit-1 + 61, Company Name a 6 6 1 American Electric Po. -9.240 0.005 0.130 0.093 {-1.610) 1.640) ( 2.032)* ( 1.520) Arizona Public Ser. -65.808 0.033 0.202 0.176 (-8.133)* 8.161)* ( 2.243)* ( 2.044)* Bay State Gas 38.412 0.019 0.176 0.119 -3.390)* 3.405)* ( 1.672) ( 1.188) Carolina Power ~3.276 0.002 0.107 0.086 («0.233) 0.256) ( 0.897) ( 0.924) Columbia Gas -28.237 0.014 0.132 0.082 -6.164)* 6.196)* ( 2.589)* ( 1.684) Commonwealth Ed. -17.151 0.009 0.204 0.207 (-2.392)* 2.417)* ( 2.551)* ( 2.718)* Consolidated Edison 20.550 0.010 0.151 0.104 -3.782)* 3.797)* ( 2.489)* ( 1.811) Delmarva P&L 19.671 0.019 0.025 -0.004 o2.825)* 2.860)* ( 0.318) (-0.052) Detroit Edison 18.190 0.009 0.186 0.153 -2.764)* 2.793)* ( 2.536)* ( 2.184)* Eastern Utilities -34.649 0.018 0.156 0.158 -5.128)* 5.160)* ( 2.221)* ( 2.296)* ENSERCH Corporation -4.908 0.003 -0.009 -0.010 (-3.798)* 3.921)* (~1.408) (-l.391) Florida Progress 53.916 0.027 0.323 0.233 -4.986)* 5.003)* ( 2.679)* ( 2.036)* 69 Table 6 (cont'd) %P,, - a:1 + BDYEAR, + 61RETURN1, + r,RE’I‘URN,,.1 + 6,, Company Name a fi 6 7 Idaho Power -42.577 0.022 0.144 0.132 (-5.016) ( 5.047)* ( 1.826) 1.768) Kentucky Utilities -2S.012 0.013 0.188 0.149 (-2.848)* ( 2.861)* (1.925)* 1.601) Louisiana General -16.857 0.009 0.013 0.008 (-7.520)* ( 7.598)* ( 0.512) 0.335) Minnesota Power -40.397 0.021 0.172 0.099 (-4.242)* ( 4.277)* ( 1.737 1.017) Montana Power 36.818 0.019 0.247 0.206 -1.696) ( 1.708) ( 1.795) 1.484) Ohio Edison 30.558 0.016 0.131 0.109 -6.764)* ( 6.809)* ( 2.605)* 2.271)* Oklahoma Power 38.716 -0.019 -0.048 0.172 ( 1.108) (-1.100) {-0.161) 0.741) Orange & Rockland -21.399 0.011 0.137 0.034 -3.933)* ( 3.970)* ( 2.253)* 0.586) Pennsylvania P&L -17.941 0.009 0.041 0.077 (-4.112)* ( 4.177)* ( 0.841) 1.673) Piedmont Natural Gas 55.712 0.028 0.106 0.090 (-3.927)* ( 3.953)* ( 0.885) 0.959) Philadelphia El. -40.156 0.020 0.089 0.094 (-9.908)* ( 9.966)* ( 1.965)* 2.197)* Potomac Electric -8.688 0.004 0.091 0.028 {-1.318) ( 1.343) ( 1.235) 0.395) Public Ser. Colorado -5.991 0.003 0.330 0.133 (-0.237) ( 0.247) ( 1.257) 0.517) 70 Table 6 (cont'd) 42,, - a, + 5,129.11, + 511mm,, + riRETURNu-1 + e,, Company Name a B 6 1 Rochester C & L -14.585 0.007 0.229 0.125 -2.l31)* ( 2.164)* ( 3.010)* ( 1.726) San Diego Gas & El. -58.009 0.029 0.099 0.178 -6.211)* ( 6.230)* (0.947) ( 1.793) Sierra Pacific -28.155 0.014 0.203 0.149 -3.894)* ( 3.931)* ( 2.522)* ( 1.946)* Southwestern Public -13.701 0.007 0.086 0.169 -0.738) ( 0.756) ( 0.546) ( 1.370) Sun Company -12.597 0.006 -0.001 0.002 (-12.404)* (12.593)* (-0.199) ( 0.364) UGI Corporation -29.207 0.015 0.012 0.003 -9.613)* ( 9.659)* ( 0.362) ( 0.094) Union Electric -34.424 0.017 0.154 0.143 -5 776)* ( 5.809)* ( 2.316)* ( 2.265)* t-statistics in parentheses * Significance at the 5% level Table 7 The Trend of Shareholder Participation and the Effect of Return on Participation Levels as a Percent of Outstanding Shares Utility Sub-Group 22,, - a, + 5,329.11, + 5,RETURN,, + riRETURN1,-1 + 5,, Company Name a fi 6 7 American Electric Po. 25.117 -0.013 0.218 0.189 ( 1.525) (-1.517) ( 1.187) ( 1.084) Arizona Public Ser. 21.812 0.011 0.145 0.119 -4.661)* ( 4.678)* ( 2.781)* ( 2.401)* Bay State Gas -4.693 0.002 0.115 0.059 (-0.772) ( 0.782) ( 2.044)* ( 1.091) Carolina Power 38.740 -0.019 0.122 0.077 ( 1.824) {-1.821) ( 0.678) ( 0.548) Columbia Gas -2.598 0.001 0.047 0.039 (-0.975) ( 0.985) ( 1.593) ( 1.376) Commonwealth Ed. 7.889 ~0.004 -0.0004 0.262 ( 0.430) (-0.418) (-0.002) ( 1.596) Consolidated Edison -9.825 0.005 0.090 0.073 (-3.272)* ( 3.280)* ( 2.670)* ( 2.304)* Delmarva P&L -15.883 0.008 0.078 -0.043 (-S.870)* ( 5.917)* ( 2.687)* (-1.541) 'Detroit Edison 7.089 -0.004 0.225 0.117 ( 0.668) (—0.655) ( 1.904)* ( 1.037) Eastern Utilities -44.188 0.022 0.239 0.185 (-6.001)* ( 6 024)* ( 3.131)* ( 2.467)* ENSERCH Corporation 2.128 -0.001 0.002 0.002 (12.644)* (-12.534)* ( 2.263)* ( 1.967)* Florida Progress 24.004 0.012 0.288 0.390 -2.237)* ( 2.243)* ( 2.405)* ( 3.424)* 72 Table 7 (cont'd) 52,, - a, + fimYEAR, + SiRETURN“ + riRETURNit-1 ‘+ .1, Company Name a 6 6 1 Idaho Power 16.929 0.009 0.062 0.067 -3.823)* ( 3.843)* ( 1.509) ( 1.710) Kentucky Utilities 3.784 -0.002 0.111 0.153 ( 0.412) (-0.405) ( 0.954) ( 1.529) Louisiana General 1.827 -0.001 0.019 0.007 ( 2.012)* (-l.967)* ( 1.885) ( 0.731) Minnesota Power 11.414 -0.006 0.102 0.083 ( 1.167) (-1.159) ( 1.001) ( 0.831) Montana Power 27.302 -0.014 -0.0007 -0.021 ( 5.519)* (-5.505)* (-0.041) {-1.299) Ohio Edison 13.752 -0.007 0.0801 0.191 ( 1.048) (-1.035) ( 0.652) ( 1.661) Oklahoma Power 53.720 -0.027 0.150 0.093 ( 2.422)* (-2.419)* ( 0.111) ( 0.662) Orange & Rockland 9.821 -0.005 0.132 0.059 ( 1.576) (-1.553) ( 2.313)* ( 1.083) Pennsylvania P&L 12.072 -0.006 0.131 0.105 ( 1.475) {-1.450) ( 1.431) ( 1.205) Philadelphia El. -7.393 0.004 0.099 0.071 {-1.733) ( 1.767) ( 2.079)* ( 1.572) Piedmont Natural Gas 24.526 0.0124 0.104 0.075 -2.876)* ( 2.897)* ( 1.442) ( 1.328) Potomac Electric 16.313 -0.008 0.0570 -0.020 ( 2.235)* (-2.212)* ( 0.700) (-0.259) Public Ser. Colorado 23.189 -0.116 0.246 0.171 ( 0.936) (-0.932) ( 0.956) ( 0.676) Rochester C & L 0.111 0.0001 0.223 0.129 ( 0.0015) ( 0.019) ( 2.615)* ( 1.585) 73 Table 7 (cont'd) 49,, - a, + 5,,th + 6,RETURN,, + r,RETURN,,_, + e,, Company Name a fi 6 7 San Diego Gas 6 E1. ~16.432 0.008 0.132 0.161 (-2.322)* ( 2.332)* 1.680) 2.149)* Sierra Pacific -9.486 0.005 0.166 0.166 (-1.198) ( 1.218) 1.885)* 1.978)* Southwestern Public 28.435 -0.014 0.112 0.099 ( 1.294) (-1.289) 0.601) 0.681) UGI Corporation 27.792 0.014 0.019 0.008 (-5.586)* ( 5.599)* 0.351) 0.153) Union Electric -8.929 0.005 0.097 0.080 (-1.490) ( 1.513) 2.158)* 1.792) t-statistics in parentheses * Significance at the 5% level Sun Company is omitted as discussed in the text. 74 Table 8 The Trend of Shareholder Participation and the Effect of Return on Participation Levels as a Percent of Shareholders Outstanding Nonutility Sub-Group Company Name 6 6 r ALCOA 0.008 -0.000 0.002 0.003 ( 0.023) -0.011) ( 1.113) ( 1.570) AMAX, Inc. 12.778 -0.006 -0.060 -0.044 ( 0.535) -0.529) (~0.517) (-0.364) Ashland Oil, Inc. -0.236 0.0001 0.079 0.072 ( -0.039) 0.045) ( 2.523)* ( 2.210)* Chase Manhattan -74.641 0.038 0.043 0.011 (-10.593)* 10.637)* ( 1.762) ( 0.484) Contel Corporation -8.457 0.004 0.003 0.002 ( -2.498)* 2.566)* ( 0.177) ( 0.097) Eaton Corporation -19.314 0.010 0.048 0.029 ( -2.135)* 2.143)* ( 1.019) ( 0.593) Equifax Inc. 0.292 —0.0001 -0.001 -0.0003 ( 2.169)* -2.083)* (-0.970) (-0.530) Equitable Resources 12.068 -0.006 0.033 0.015 ( 4.835)* -4.813)* ( 3.822)* ( 1.760) General Mills -0.182 0.0001 0.001 0.001 ( -1.360) 1.433) ( 1.370) ( 0.850) Gillette Company -0.027 0.0000 -0.0004 -0.0002 ( -0.172) 0.229) (v0.552) (-0.249) Johnson Controls -l.029 0.001 0.006 0.013 ( -0.665) 0.676) ( 0.800) ( 1.789) 75 Table 8 (cont'd) 42,, - a, + 5,525.2, + 6,RETURN,, + r,RETURNL,,_, + e,, Company Name a 6 6 r Mead Corporation -2.567 0.001 -0.006 -0.005 ( -2.921)* ( 2.928)* (-1.331) (-1.072) Minn. Mining 6 Man. -1.802 0.001 -0.002 -0.001 ( -8.508)* ( 8.565)* (-2.205)* (-0.948) Mobil Corporation -1.338 0.001 -0.001 -0.0002 ( -7.443)* ( 7.503)* (-0.643) (-0.204) NCNB Corporation -17.836 0.009 -0.010 -0.023 ( -3.508)* ( 3.544)* (-0.391) {-0.859) Panhandle Eastern -22.639 0.012 -0.018 -0.039 ( -3.358)* ( 3.392)* (-0.515) (-1.076) PPG Industries -0.463 0.0002 -0.003 -0.002 ( -1.661) ( 1.755) (-1.955)* (-1.587) Textron, Inc. 0.791 -0.0004 0.001 -0.003 ( 1.263) ( -1.240) ( 0.407) (-0.810) Thomas 6 Betts -1.135 0.001 -0.001 -0.001 ( -7.282)* ( 7.351)* (~l.065) (-0.855) Union Pacific —0.130 0.0001 -0.001 -0.0005 ( -0.738) ( 0.754) (~1.694) (-l.103) Upjohn Company -1.880 0.001 0.003 0.004 ( -4.221)* ( 4.308)* ( 1.259) ( 1.758) t-statistics in parentheses * Significance at the 5% level Kerr-McGee and Ralston Purina are omitted as discussed in the text. 76 Table 9 The Trend of Shareholder Participation and the Effect of Return on Participation Levels as a Percent of Eligible Shareholders Nonutility Sub-Group 42,, - a, + 5,,Y2AR, + 6,RETURN,, + r,RETURN,,-, + 6,, Company Name a B 6 r ALCOA -22.784 0 012 0.033 0.055 ( -5.503)* 5.532)* ( 1.538) ( 2.466)* AMAX, Inc. -47.771 0.024 -0.021 -0.028 ( -6.815)* 6.852)* (-0.588) (-0.734) Ashland Oil, Inc. -27.069 0.014 0.052 0.030' ( -6.124)* 6.172)* ( 2.263)* ( 1.261) Chase Manhattan -17.777 0.009 -0.006 -0.008 ( -l.716) 1.730) (-0.179) (-0.229) Contel Corporation 3.872 -0.002 -0.015 -0.034 ( 0.933) -0.868) (-0.679) (-1.528) Eaton Corporation -16.448 0.008 0.036 0.021 ( -4.875)* 4.917)* ( 2.088)* ( 1.148) Equifax Inc. -14.729 0.007 -0.014 -0.005 ( -8.938)* 9.005)* (-l.922)* (-0.677) Equitable Resourses -22.823 0.012 0.013 0.033 ( -1.6l2) 1.628) ( 0.273) ( 0.701) General Mills -27.234 0.014 -0.040 -0.033 ( -4.693)* 4.728)* (-l.329) (-1.069) Gillette Company ~13.999 0.007 0.005 0.002 (-22.787)* 22.957)* ( 1.496) ( 0.707) Johnson Controls -38.960 0.020 0.014 0.114 ( -3.681)* 3.713)* ( 0.298) ( 2.307)* Kerr-McGee Corp. -24.662 0.012 0.003 -0.007 (~13.302)* ( 13.362)* ( 0.296) (-0.737) 77 Table 9 (cont'd) 22,, - a, + 5,,Y2AR, + 6,RETURN,, + r,RETURN,,-, + e,, Company Name a fi 6 r Mead Corporation -17 336 0.009 0.017 0.018 ( -3.770)* ( 3.812)* ( 0.725) ( 0.710) Minn. Mining 6 Man. -5.138 0.003 -0.007 -0.0003 ( -5.860)* ( 6.006)* (-1.497) (-0.057) Mobil Corporation -18.682 0.009 -0.014 -0.011 ( ~9.442)* ( 9.486)* (-l.320) (-0.973) NCNB Corporation 0.769 -0.0002 0.018 0.026 ( 0.115) ( -0.074) ( 0.509) ( 0.715) Panhandle Eastern -20.681 0.011 0.034 0.026 ( -6.308)* ( 6.366)* ( 2.020)* ( 1.500) PPG Industries -14.252 0.007 ~0.027 -0.007 ( -3.771)* ( 3.813)* (-1.404) (-0.366) Ralston Purina -30.858 0.016 0.014 0.026 (-10.488)* ( 10.564)* ( 1.051) ( l.910)* Textron, Inc. -16.301 0.008 0.013 -0.015 ( -3.605)* ( 3.631) ( 0.533) (-0.603) Thomas 6 Betts -26.224 0.013 -0.014 -0.007 (-11.237)* ( 11.302)* (-1.143) (-0.534) Union Pacific -11.605 0.006 -0.055 -0.025 ( -1.754) ( 1.786) (-2.439)* (-1.396) Upjohn Company -27.019 0.014 0.015 0.019 ( -7.775)* ( 7.847)* ( 0.851) ( 1.029) t-statistics in parentheses * Significance at the 5% level 78 are positive, with 73.9% of the runs positive and significant at the 5% level. The contemporaneously and lagged return variable coefficients are mostly positive suggesting a direct relationship between return and participation. Over 90% of the firms had positive coefficients while 46.9% where positive and significant for contemporaneous returns, and 25% where positive and significant for a one year lag return. Table 7 shows 93.5% and 90.3% positive coefficients with 41.9% and 22.6% significant for contemporaneous and lagged returns respectively. Table 8 and 9, which deal with the nonutility subgroup, has somewhat weaker results. A little less than 50% of the contemporaneous and lagged return variables had positive coefficients while only 5-13% were both positive and significant. The results are contradictory to Pettway and Malone's (1973) result which suggested return does not affect participation rates. These results support the notion that _ participation rates in DRPs are affected positively by the stock's return. 5.3b Wis To further develop evidence in support of Proposition 3, the model is run on a cross sectional sample of the firms. The data set is pooled in the following manner: 79 %PR Year, %Pn_ Year, %P, = %P,, Year, = Year, spa, Year, Return,, Return,,.1 Returnz, Return,,.1 Return, -= Return3, ‘ Return,.1 = Return,,_1 Return,, Return,,.1 Equation 9 will be run with the pooled data for each subgroup, utility and nonutility. %P, = a + 8Year, + 6Return, + rReturn,_1 + e, (9) Again, participation will be represented by percent of eligible shareholders and by percent of outstanding shares, thus two runs of equation 9 will be run for the utility subgroup and the nonutility subgroup. When analyzing the results of the regression, autocorrelation was found and needed to be Corrected. The Hildreth-Lu method for correcting autocorrelation was used. The results of the four regressions are shown in Tables 10 and 11. For the utility subgroup, see Table 10, a positive and significant coefficient was found for the trend variable as expected. Of interest is the significant and positive coefficient on the lagged return variable in the regression using percent of eligible shareholders for participation, and the positive and significant coefficient for the 80 Table 10 Effects of Stock Return on Participation Rates Pooled Sample - Utilities %P, = a + BYear, + 6Return, + rReturn,.1 + e, DEPENDENT a B 6 r VARIABLE Participation as % Of -34.226 0.017 0.014 0.029 eligible (—16.470)* (16.589)* (1.195) (2.668)* shareholders Participation . as % Of -10.560 0.005 0.021 0.017 outstanding (- 4.980)* ( 5.044)* (1.857)* (1.535)# shares * Significance at the 5% level # Significance at the 10% level t-statistics in parentheses 81 contemporaneous return variable using percent of outstanding shares for participation. In fact, contemporaneous return is positive (participation as percent of eligible shareholders) and lagged return (participation as percent of shares outstanding) is positive and significant at the 10% level. These results support Proposition 3. It appears that as the return of a stock increases, participation rates also increase. Of course, investors would have to expect that the prior and current returns are going to continue in the future. For the nonutility subgroup, see Table 11, it can be seen that only the trend variable is positive and significant. This result supports Baker and Meeks' (1990) observation of an upward trend in participation. The return variables were not significant or positive. No support for Proposition 3 can be given with the nonutility subgroup. 5.4 Methodology and Results of Proposition 4 2129295111254 The percent participation in a firm's dividend reinvestment plan with respect to a) the number of shareholders and b) the number of total shares will have a negative effect on the volume of trade in the firm's stock. To test this proposition, the pooled sample of firms will be tested using various forms of equation 10. v, = a + sv,, + 52?, + e, (10) 82 Table 11 Effects of Stock Return on Participation Rates Pooled Sample - Nonutilities %P, = a + BYear, + 6Return, + rReturn,.1 + e, DEPENDENT a B 6 r VARIABLE Participation as % Of -27.585 0.014 -0.014 0.005 eligible (~13.939)* (14.031)* (-l.582) (0.511) shareholders ' Participation as % Of - 6.700 0.003 -0.002 0.001 outstanding (- 2.658)* ( 2.680)* (-0.l47) (0.071) shares * Significance at the 5% level t-statistics in parentheses 83 where %Plt vmt ' Vlt %Pzr. vmt V2: %Pt 3 %PBt vmt ' mt Vt = Vat. %Pnt vat, vnt As was discussed earlier, the equation will be run once with participation rates in terms of eligible shareholders and again with participation rates in terms of outstanding shares. For the volume variables, the natural log of the volumes was used. Also, autocorrelation was a problem and was corrected for using the Hildreth-Lu method. As discussed in Chapter 3, if shareholders are buy-hold investors, then the coefficient on the participation term should be negative. Stated formally, Hypothesis 5-4 H,: Participation in DRPs has no effect on firm trading volume. H,: If participation in DRPs is large enough, the volume of trading should decrease. Table 12 shows the regression results of the utility subgroup and Table 13 shows the regression results of the nonutility subgroup. As can be seen, the sign of the participation variable is opposite to what is expected in three of the four trials. In the trial in which the 84 Table 12 Effects of Participation Rates on Trading Volume Pooled Sample - Utilities V,=a + BVm-l- 6%P,+ 6, DEPENDENT a B 5* VARIABLE NATURAL LOG -10.303 1.240 0.647 OF FIRM (- 9.217)* (10.691)* (0.836) TRADING VOLUME a B 6# NATURAL LOG -10.506 1.250 1.855 0F FIRM (-10.886)* (13.834)* (2.314)* TRADING VOLUME 6* - DRP participation as a percent of eligible shareholders. 6# - DRP participation as a percent of outstanding shares. * Significance at the 5% level t-statistic in parentheses. 85 Table 13 Effects of Participation Rates on Trading Volume Pooled Sample - Nonutilities DEPENDENT a B 6* VARIABLE NATURAL LOG -9.364 1.264 -1.055 OF FIRM (-8.985)* (11.717)* (-1.215) TRADING VOLUME a B 6# NATURAL LOG -8.018 1.108 0.556 OF FIRM (-7.746)* (11.505)* (0.564) TRADING VOLUME 6* - DRP participation as a percent of eligible shareholders. 6# - DRP participation as a percent of outstanding shares. * Significance at the 5% level t—statistic in parentheses. 86 expected sign is found, the coefficient is not significant even at the 10% level. Support for Proposition 4 is not found. For the utility subgroup, the coefficient of the participation as a percent of outstanding shares is significant but positive. While this result does not support Proposition 4, it may suggest that large institutional investors, that often hold big blocks of stock, increase the firm's volume of trade. Institutional investors may purchase stocks, take advantage of reinvesting dividends at low or no brokerage costs and/or discounts if offered, and then sell at a later date for profits. Thus if large block shareholders are not buy and hold investors, the potential for increased trading volume arises as participation in DRPs increases. Again, as in Proposition 2, an in depth look at the exact make up of participants would need to be performed in order to find support for the above supposition. 5.5 Conclusion DRPs have been shown to be attractive to investors because of the low cost of reinvesting, the discount from market price offered by some plans, and in 1982 - 1985, the tax deferment of dividends that were reinvested. Also, Chapter 5 provided some evidence that participation tends to be positively related to stock return, contemporaneous and lagged. As discussed in the prior section, it appears that 87 participation rates have little or no significant effect on trading volume. The survey results were used to understand the characteristics of DRPs and to ascertain when and why investors might join the plans. The next two chapters will focus more on the potential effects dividend reinvestment plan usage by a firm may have. Chapter 6 explores the investing in portfolios made up of stock's with and without DRPs, while Chapter 7 deals with the ex-dividend day phenomena. The results of these two chapters can be used to assess if considering the use of a DRP in the dividend decision is of value to management. CHAPTER 6 RETURN OF PORTFOLIOS CONSISTING OF STOCKS WITH DRPs To assess the impact DRPs may have on portfolio return, this study compares returns of portfolios containing DRPs with those that do not contain DRPs. If the DRP adds some value to the firm's stock, portfolios of stock that use DRPs may outperform similar portfolios of stock that do not contain DRPs. As discussed below, this study will use only stocks in the utility industry to eliminate any potential industry effects. 6.1 Prior Work on Portfolios of Stocks with DRPs Fetherston and Hill (1990) found that a stock portfolio comprised of 505 firms using DRPs outperformed the market over the period of 1968-1987. Realizing the cost prohibiting factor of investing in 505 stocks, they randomly selected four portfolios containing 20 stocks each. They also formed a portfolio with the 20 worst performing DRP stocks. Their results show that all four randomly generated portfolios outperformed the market: however, the portfolio of lowest performing DRP stocks miserably underperformed the market. 88 89 The cause of the better than market performance of the random portfolios is unknown. It could be that one or more industries are driving the result. Fetherston and Hill did separate the sample into industries: however, there is no comparison of these portfolio returns with portfolio returns comprised of other firms in the same industry group that do not have DRPs. Therefore one is unable to distinguish if the DRP is driving the better than market return or if the industry group itself is driving the return. In other words, it may be that stocks in these industries would have had better than market return whether or not the particular firm used the DRP. In Table 4 of their article, they listed the 20 best and twenty worst performing industries in terms of average return. To better understand how these industries' returns compared to the market return on a risk adjusted basis, a simple average return per standard deviation was calculated?. The results are reported in descending order in Table 14. Fifteen industries performed better than the market on a risk-adjusted basis. Table 15 reports the industries in descending order based upon compound return. It is noted that there are only eight industries that performed worse than the market based on compound returns "The industries and numbers used in the calculations are taken directly from Table 4 of Fetherston and Hill (1990, ‘p.35) article. 'The numbers used may be subject to calculation error as the published equations were incorrect. It is unclear as to whether the equations published are typos or if they were actually used. 90 Table 14 Fetherston and Hill (1990) Table 4 - Industry Return SORTED BY AVERAGE RETURN PER STANDARD DEVIATION Industry Compound Average Std . Ave Ret / Group Return Return Dev . Std Beverages (nonalc) 16.0 17.2 16.7 1.0299 Telephone 11.5 12.3 12.6 0.9762 Food Processors 15.9 17.5 20.0 0.8750 Tv/Radio/Motion Pictures 12.0 20.0 24.3 0.8230 Railroads 14.8 16.9 20.6 0.8204 Printing/Pub/News 14.5 16.4 20.6 0.7961 Banks/Money Center 10.2 11.3 14.9 0.7584 Containers (paper) 15.8 18.9 25.7 0.7354 Machinery/Handtool 11.5 12.9 1737' 0.7288 Pipeline (oil/gas) 11.4 12.7 17.5 0.7257 Electric Utilities 10.7 12.0 16.6 0.7229 Non-bank Financial 13.9 16.4 23.8 0.6891 .Auto/Parts/Rental 11.1 19.5 28.7 0.6794 Banks/Regional 11.0 13.0 19.7 0.6599 Variety/Dept.stores 16.4 20.4 31:7 0.6435 Market 9.3 10.6 16.8 0.6304 Conglomerates 9.8 11.2 17.9 0.6257 Insurance 9.0 10.7 18.7 0.5722 Drugs/Retail 18.0 25.1 44.1 0.5692 Machinery 9.8 12.0 21.5 .0.5581 Construction Equipment 16.7 23.4 -42.3 0.5532 Office/Computer 9.8 12.2 22.2 0.5495 Health/Production & Serx 13.9 20.4 37.5 0.5440 Advertising Agencies 13.0 16.9> 31.9 0.5298 Waste/Service 13.0 19.4 36.8 0.5272 Oil Producers 11.9 16.3 32.5 0.5015 Real Estate 9.4 12.6 25.5 0.4941 Steel 9.9 12.6 25.8 0.4884 Apparel/Shoes 8.8 12.2 26.2 0.4656 91 Table 14 (con't) Auto/Rubber 8.8 11.6 25.1 0.4622 Freight (air/ground) 7.4 9.9 22.4 0.4420 Food Stores 15.2 17.3 42.2 0.4100 Plastic/Rubber 10.2 16.0 39.1 0.4092 Aluminum 6.8 9.4 24.4 0.3852 Air Lines 7.0 10.6 28.9 0.3668 Auto/Manufacturing 10.6 18.4 51.6 0.3566 Photography 5.6 10.5 31.2 0.3365 Coal 2.7 8.6 35.8 0.2402 Note: The numbers for compound return, average return and standard deviation come directly from Table 4 of Fetherston and Hill (1990, p.35) 92 Table 15 Fetherston and Hill (1990) Table 4 - Industry Return SORTED BY COMPOUND RETURN Industry Compound Average Std . Ave Ret/ Group Return Return Dev . STD Tobacco 18.2 19.1 14.0 1.3643 Drugs/Retail 18.0 25.1 44.1 0.5692 Construction Equipment 16x7 23u4 42.3 0.5532 Variety/Dept.Stores 16.4 20.4 31(7 0.6435 Beverages (nonalc) 16.0 17.2 16.7 1.0299 Food Processors 15.9 17.5 20.0 0.8750 Containers (paper) 15.8 18.9 25.7 0.7354 Food Stores 15.2 17.3 42.2 0.4100 Railroads 14.8 16.9 20.6 0.8204 Printing/Pub/News 14.5 16.4 20.6 0.7961 Health/Production & Serx 13.9 20.4 37.5 0.5440 Non-bank Financial 13.9 16.4 23.8 0.6891 Advertising Agencies 13.0 16.91 31.9 0.5298 Waste/Service 13.0 19.4 36.8 0.5272 TV/Radio/Motion Pictures 12.0 20.0 24.3 0.8230 Oil Producers 11.9 16.3 32.5 0.5015 Machinery/Handtool 11.5 12.9 17.7 0.7288 Telephone 11.5 12.3 12.6 0.9762 Pipeline (oil/gas) 11.4 12.7 17.5 0.7257 Auto/Parts/Rental 11.1 19.5 28.7 0.6794 Banks/Regional 11.0 13.0 19.7 0.6599 Electric Utilities 10.7 12.0 16.6 0.7229 ' Auto/Manufacturing 10.6 18.4 51.6 0.3566 Plastic/Rubber 10.2 16.0 739.1 0.4092 Banks/Money center 10.2 11.3 14.9 0.7584 Steel 9.9 12.6 25.8 0.4884 Machinery 9.8 12.0 21.5 0.5581 Conglomerates 9.8 11.2 17.9 0.6257 Office/Computer 9.8 12.2 22.2 0.5495 Real Estate 9.4 12.6 25.5 0.4941 Market 9.3 10.6 16.8 0.6304 93 Table 15 (con't) Insurance 9.0 10.7 18.7 0.5722 Apparel/Shoes 8.8 12.2 26.2 0.4656 Auto/Rubber 8.8 11.6 25.1 0.4622 Freight (air/ground) 7.4 9.9 22.4 0.4420 Air Lines - 7.0 10.6 28.9 0.3668 Aluminum 6.8 9.4 24.4 0.3852 Photography 5.6 10.5 31.2 0.3365 Coal 2.7 8.6 35.8 0.2402 Note: The numbers for compound return, average return and standard deviation come directly from Table 4 of Fetherston and Hill (1990, p.35) 94 during the time period. It seems plausible that the industries with better than market performance may be driving the result and not the DRP'.g This study addresses the potential impact of DRPs on portfolios of one industry group, utilities. This group is chosen for two reasons. One, two of the industry groups that have higher average return per risk than the market over the time period are utilities - telephone, which has the third highest risk adjusted return and electric utilities. Additionally, as noted in the survey section, utilities are large users of dividend reinvestment plans, which allows the number of firms used in the sample to be of adequate size for testing. Thus, for the utility industry this work will attempt to determine whether better than market return is caused by the firms use of DRPs or some other industry factor. 6.2 Utility Stock Portfolio Returns To better understand the effect DRPs may have on the returns of portfolios, the utility industry will be separated into portfolios of firms using the DRP and firms that do not. A return on these subgroups is calculated and compared. Also, a comparison to the market return is made. 9Industries with better than market returns would have to more than offset the industries with lower than market returns for this assertion to be true. 95 6.2a Data The first dividend reinvestment plan was initiated in 1968, thus only a few companies were using the plan at this time. To allow time for other firms to begin using the plan, the return data does not begin until 1972 and runs until 1989. Also, information on which firms had DRPs as reported in Stangarg_ang_figg;§_gutlggk was only accessible beginning in 1972. The sample consists of 107 utility firms that have a complete stock return series for 1972 - 1989. The return data for the individual firms and the market were obtained from the §B§£ monthly data tapes. The monthly returns were then annualized by 12 km =." (1'*}9u) ’ 1- (3) 3‘1 where kut== the monthly return for firm i or the market in month j, and kn = the annual return for firm i or the market in year t. 6.2b Mstnedglesx_and_H222tb2§i§ For each year, the sample was split into firms that have DRPs, firms that have discount DRPs (DDRPs) and firms without dividend reinvestment plans. The information as to whether the firm had a plan and if it had a discount was obtained from Ihe_SLandard_and_2222§_gutlggk. which publishes a list of firms with DRPs and any characteristics 96 pertaining to those plans about once a year. The sample was not split into firms with DDRPs until 1977. In 1975 AT&T initiated the first discount; however, the §tgn§g:g_ang EQQ:§_Qut19gk did not report discount information until 1977. The Stgndg:§_gnfi_£gg:§_ggtlggk unfortunately does not publish the list of firms with DRPs at the same time each year. When grouping the firms, if a firm was reported to have a plan or a discount at the time the Stgnggrg_gng_figgzs Outlook published the information, it was assumed that the firm had these characteristics throughout the full year. The number of firms in each category varies with each year.10 Once the firms were separated into portfolios by usage or nonusage of DRPs, an average return and standard deviation of each portfolio for each year was calculated using the following equations. n . Average Return = ARtq = 2 km! (10) i=1 n Standard Deviation = STDtq = 1 n [ — 2 (km; - 412..)Z 1" (11) n-l i=1 where k1,;q = return on firm i in year t in group q 10The number of firms not using the DRP became less than ten after 1980 and as low as five in 1986. 97 q - 1, 2, or 3 with 1 = no DRP, 2 = DRP, 3 = DDRP portfolios. Once yearly returns were calculated, holding period returns and standard deviation of return for 1972 - 1989 and 1977 - 1989 were calculated using the following equations. Holding Period Average Return = HPAVq== AR“! (12) 1 m "D45 t Standard Deviation of the Portfolio = STDq== l m [— 2 (ARWI - Hpavqf ]* (13) m-l t=1 where m = the number of years in the holding period, and q equals 1,2,3,4 with 1-3 defined as before and 4 equal to the market portfolio. Based on Fetherston and Hill's (1990) conclusion that the DRP adds value to a portfolio, it is expected that the portfolio of utilities that have dividend reinvestment plans should outperform the portfolio of utilities without the plan on a risk adjusted basis. Stated formally Hypothesis 6-1 H5: There is no difference in the risk adjusted return of a portfolio of utility stocks with DRP and one without. 98 fig: Risk adjusted returns of a portfolio of utility stocks with DRPs are higher than risk adjusted returns of a portfolio without DRPs. If the DRP causes portfolio return to be greater than the market, the portfolio without the DRP should underperform or perform equally as well as the market. If the no DRP portfolio also outperforms the market, then one can not conclude that the DRP causes better than market return. Therefore the following hypothesis is offered. Hypothesis 6-2 Ho: There is no difference in the risk adjusted return of a portfolio of stocks containing no DRP, containing DRPs and the market. H.: The risk adjusted returns of the DRP portfolio is greater than the market portfolio which is greater than or equal to the no DRP portfolio. 6.3 Results This section will discuss the results of the testing of the proposed hypotheses. 6-3a W Table 16 displays the results of the portfolio holding period returns. The arithmetic mean return of the no DRP and DRP portfolios over the holding period of 1972 - 1989 99 Table 16 Portfolio Returns of Utility Stocks Holding Period Returns Arithmetic Return Geometric Return 1972-1989 1977-1989 1972-1989 1977-1989 No DRP Portfolio 0.16501 0.185712 0.1437 0.17019 Sharpe-R 0.39246 0.52854 - - .----------.b--------------—---‘-------------—-----1 DRP Portfolio 0.16146 0.17961 0.14524 0.17285 _§harpe4R_ 0.43526___0:7616 _:;________:____ DDRP Portfolio - 0.16595 - 0.15782 399523213 ...... : ...... 9:33.89}. ---_:__-______: ..... Market Return 0.132924 0.15496 0.11836 0.10490 ‘Sharpe-R 0.30803 0.5527 - - Note: Sharpe-R is the Sharpe Ratio used to measure risk-adjusted portfolio return. 100 are both greater than the market's average return. The arithmetic mean return of the no DRP portfolio however, is slightly larger than the return on the DRP portfolio which is contrary to expectations. But when comparing the mean returns on a risk adjusted basis, using the Sharpe ratio“, results supporting Hypothesis 6-1 are found. The risk/return ratio on the DRP portfolio is larger than the risk/return ratio on the no DRP portfolio, 0.43526 vs. 0.39246. The risk/return ratio of the no DRP portfolio is higher than market's risk/return ratio. These risk adjusted arithmetic mean results do not allow one to reject the null of Hypothesis 6-2. I When the holding period of 1977-1989 is examined, the portfolios containing DRPs and DDRPs have higher returns than the market return, but the no DRP portfolio also has a higher return than the two DRP portfolios, which is contrary to expectations. However, when looked at on a risk/return basis, the results reject the null Hypotheses 6-1 and 6-2. The DRP portfolio's risk/return ratio is greater than the market's risk/return ratio which is greater than the no DRP portfolio risk/return ratio. The portfolio consisting of firms with DDRPs behaves as expected in the alternative of Hypothesis 6-1; its risk/return ratio is larger than the portfolio of firms 11The Sharpe Ratio is calculated by subtracting the risk free rate, proxied by 3mo T-bill returns, from the portfolio return. This difference is then divided by the portfolio's standard deviation. 101 without DRPs risk/return ratio. This result also rejects the null Hypothesis of 6-2 which contends that portfolios containing DRPs will not outperform the market. It is interesting to note that the discount DRP portfolio underperforms the portfolio containing stocks with DRPs but not offering discounts. 6.310 W The results for the Geometric return over the two holding periods studied are also shown in Table 16. These results show that for the 1972-1989 holding period the no DRP portfolio has a slightly lower return than the DRP portfolio, thus supporting the contention that the dividend, reinvestment plan adds value to the portfolio return. Both of the above discussed portfolios have a greater geometric return than the market, thus the results do not fully support the assertion that the DRP is causing the better than market return. The utility group as a whole seems to outperform the market over this time period. The 1977-1989 time period provides results similar to the longer time period. The DRP portfolio outperforms the no DRP portfolio and both of these portfolios outperform the market. As with the arithmetic mean return results, the DDRP portfolio while performing better than the market, underperforms the no DRP and DRP portfolios. Whereas the arithmetic risk/return ratio results reject the null of Hypothesis 6-1 and 6-2 in most cases, the /”‘ 102 geometric mean return results only reject the null of Hypothesis 6-1. The results presented in Table 16, arithmetic and geometric holding period returns, are also graphically presented in Figure 4, Figure 5 and Figure 6. 6-3C IDQIELQQQL_X§§I_B§§BLE§ To discern whether the results of the holding period returns are driven by one abhorrent time period, yearly arithmetic returns are examined. Figure 7 graphically shows the yearly returns for each of the four portfolios, while Table 17 reports the actual returns and Sharpe ratios of the portfolios. A small number of years, for example 1975 and 1988, show a greater divergence of return between utility portfolios and the market portfolio. However, it does not appear that any one particular time period is driving the holding period return results. 6.3d WW While at first glance the holding period risk adjusted return of the portfolios support expectations as written in the alternative of Hypothesis 6-1 and 6-2, a test to determine whether there is statistical significance between the mean returns of the portfolios must be conducted. Each portfolio is tested against the other portfolio within the same holding period. The hypothesis formally stated is 103 0.5 0. 45 0.4 0. 35 0.3 0.25 0.2 0.15 00' 0.05 RETURN STD SHARPE m no DRP DRP MARKET Figure 4: ARITHMETIC MEAN PORTFOLIO RETURNS - 1972—1989. This figure presents the arithmetic mean return, standard deviation of returns and risk adjusted mean returns of utility stock portfolios consisting of no DRPs, DRPs and the market over the holding period 1972-1989. 104 RETURN STD SHARP! Macaw -DRP -DDRP 82mm Figure 5: ARITHMETIC MEAN PORTFOLIO RETURNS - 1977-1989. This figure presents the arithmetic mean return, standard deviation of returns and risk adjusted mean returns of utility stock portfolios consisting of no DRPs, DRPs, DDRPs and the market over the holding period 1977-1989. 105 0J9 0J8 017 0J6 015 014 0J3 0J2 0J1 0A 009 008 007 006 005 004 003 002 001 1972—1989 1977-1989 m NO DRP DRP W/A DDRP 82 MARKET Figure 6: HOLDING PERIOD PORTFOLIO RETURNS - GEOMETRIC RETURN. This figure presents the geometric holding period returns of utility stock portfolios containing no DRPs, DRPs, DDRPs and the market over two different holding periods, 1972-1989 and 1977-1989. 106 .mwmalmhmd oowuom mafia may Ho>o umxuma oz» can ammoo .nmmo .mmmo o: no mswvuwucoo mOfiHOHHHom xooum auwawus HON mcusuou OHHOHHHOQ wauomh one musmmoum musmwu mace $373.: 1 may—3mm Sachem?“ 95224 ”p 6.33m PU¥1(I 4 ERGO 0 ...—KO ... 010 07. D ONO P 500 P DUO P MUG P P ”a P 050 P BBQ P “hm P ”NO P — UUO P — 000 P — V00 P — N00 P _ 000 P — “NO P — 050 P _ th P — “NO P b p p p b p p p n ”00' 1. “.01 107 ~mm~... m-~..~ ~m...~... .....mm. ..--... $3.... ~23... $2 $2... 4.3.... ~33... 33.... ~38... 0.8."... 8.3... 2.2 ~m~..... ~88..- 32.....- 3.8..- 82.....- 88...... $31..- ...m. :2... o..m¢.~ .32... .53.. .......~... 8.5.... 3.3... ...... .32.... 9.8.... «3...... «~80. 822.. 3.8... .32.... 32 83... 5.8... ~m$~... mm-md ...m..~... 33.....- ~28... 2.3 o.-... 833... 3...... 32...... “.33... $8.... $32.. 22 33... $8.... $3.”... 83.... 238... 88.... ~23... $2 32...... ~83... 2...... ..mm..~... 2...... 83.... .23... ...... «3m... .88... ~82... 3mm... 23.... ..-.....- .23... ....m. ....~... ~33..- .....~..... .28.... 8.2... .~..~.....- .88... m8. -.o.o .~mwm.u ......o- m.m~m.- m.m~o.ou ~.mm..- mm.~o.on ~.m. ameo.ou .o¢~xd m¢~mo.o .m..mc ~.mmo.o ¢o.mto ~m..o.o ..m. :8... m~w¢o~ 308... .33.... .32... m... .m.~.o .mm~.. oo~mm.o mommm. aom~¢.o m.m. ......~...u mm-m.~- «2.2...- .....m...~- ..-.~...u :2 83...: m....¢m.~- 8.3...- 8.....- .33...- m8. 3:... $3.... 3.3.... $82. 38.... ~.m. com: owuoazuflud Ham» Hfivwm 0.“va OHHOHHHOA Ofiumm OfiHOHuHOA OHHMM OHHOHHHOQ umxumz maumnm 9.8 waumam ...... 2.02m ...... oz mom. . ~.m. mausumm oaHOHuuom muflawuo h H OHQMB 108 Hypothesis 6-3 H5: There is no statistically significant difference between the mean return of the different stock portfolios. H.: The mean return of the stock portfolio with DRPs is larger than the market mean return which is larger than or equal to the mean return of the stock portfolio containing no DRP. To test Hypothesis 6-3, it is assumed that each of the portfolios come from a normal distribution and the variances of the portfolios are equal. The t-test derived from this assumption is very robust when the number of observations in the samples are close to being equal. The equation used to test Hypothesis 6-3 is HPAVq - HPAV, - o T = (14) S[1/nq + 1/n,]" where HPAV‘I and HPAV, equal the mean return of portfolio q and r, with q = 1,2,3 and r = 1,2,3 where 1,2,3 have the same meaning as before. r n q s = {[(n. - ms". + (n. - 1)sz.1/ (n. + n. - 211*(15) 109 where S - pooled standard deviation of the portfolios with qu and $2, equaling the sample variances of the portfolios q and r. Again q a r. As one can see from the results in Table 18, none of the mean returns between any of the portfolios is significantly different. The null of Hypothesis 6-3 cannot be rejected and the mild support found for Hypothesis 6-1 and 6-2 is not statistically significant. Therefore, this result suggests that there is no significant difference in returns of these portfolios. One can only conclude that for this set of utility stocks, it appears that the use of the DRP plays no part in causing a better than market return. The above test assumes that the variances of the portfolios are equal, therefore a pooled standard deviation was used. If the assumption is relaxed and the variances are not assumed equal, we can test to see whether there is a significant difference in the variances. A difference in the variances may show that while average return is the same . for all portfolios, an investor may be able to obtain the return for a lower risk with one portfolio than another. To have a positive effect, the presence of the DRP would, in this case, need to cause a lower variance of return for the portfolio(s) that contain DRPs. Stated formally, the test for unequal variances is 110 mOHHOuuuom 03» man coozvon :usumu some cw mocmuomuflo on» no pump on» Ho“ afiumflumuqu map ma Hogan: aouuon wee .c0auma>oo oumocmum omHoom may uuaomoumou Hones: no» one «muoz I mo¢m~.o Hoooo.o ~¢¢n¢.c mmom¢.o I mmmm..c mmeH.o emomH.c Hmmma.o umxumz IIIquIII-IIIIHIIII:ImmmmmmmI-IIIAHIIIIIIIIH ..... .mmwwwmmmII .. I 2.3.... I I 98.. IIIIanIIIIIIInIIII. ..... nIIIIIIIIu. ....... .wwwwumI.mmmwwwmmmII I I I I omomo.o mun llllllll lllllllllljrlllllllll 11111111 111111111 .lllllllllll fl mmm.m.o am.mm.o emmHH.o .. I i .. OHHOHuHom ommm..o ..mo..o Hmmma.o I I mmo oz mmeInumH mmmHIsan mmmHIhhm. mmmHImhmH mmmHImhmH umxumz ofinoMuuom oflnofifiom owaofiflom ounaowuuom Axon mmo man 55 oz mausumm OHHOHHHom comzpom noncommuuwo HMOflumwumum ma manna 111 Hypothesis 6-4 Ho: There is no difference between the variance of the portfolios that contain DRP versus those that do not contain DRPs or the market portfolio. H.: The portfolios containing DRPs have a lower variance of returns than the market portfolio or 'the portfolio that contains no DRPs. To test Hypothesis 6-4, the following F-test will be used. F = ' (16) Again, r a q. The F-value for 5% significance with degrees of freedom 17 and 17 is between 2.29 and 2.23.12 This F-value is used for the comparison of F-statistics for the portfolios over the 1972-1989 time frame. The F-value for 5% significance with degrees of freedom 12 and 12 is 2.69. This F-value is used for comparison of F-statistics for the portfolios in the 1977-1989 time period. The results of the F-test are reported in Table 19. As can be seen from Table 19, none of the F-statistics is significant and thus the null of Hypothesis 6-4 cannot be rejected. It appears from the results of Tables 18 and 19 12The F-values come from §Lé£i§LiQ§L_M§§hQ§§, Snedecor (1946, pp.222-225). 112 ..m>m. mnemoeuecmem «m was an menu. as.» mam.-..m. may no. mm.~ Una mach“ QEHH mmmalwhma ecu HOH m~.N flaw aN.N :mm3vmn mfl o=Hm>Im one .mOfiaOuuuom oz» can somzuon oocmwum> ca mocmumuuflo ecu «0 pump can no“ mOflpmHuwumIm can own mucosa: one "muoz I vmmmH.H H¢hmo.H memH.H ommmm.H uoxumz IIIInHIII-IIIIuIIIIaImmmMMHWIIIIIAHIIIIIIIIH IIIII .mwwwwmmmII .HMQQ IIIIanIIIIIIIHIIII:IIIIHIIIIIIIIMHIIIIIwwwmmumI MmmwwwmmmII an Immwmmmmu.-dmwwmumu:ummmmmnmu----JHIIIIIIIIH ..... .mmwwwmmMII a0 02 mmmfilhhma ammfilhhma mmmalhnma awaalmhmfi mmmalwhaa nexus: oonuuom OflHOHuuom ofiofitom owHOHDHom man—Q an man min 02 mmocmwum> OHHOHDHOA cmmzumm moocououuflo Havaumflumpm m. magma 113 that for the utility industry, there is no advantage to investing in utility firms that have DRP or DDRPs over the time frame studied. 6.4 Conclusions To summarize, portfolios of utility stocks were separated into three different portfolios, those containing DRPs with no discounts, those containing DDRPs and those containing no DRPs. When comparing the holding period returns of these portfolios, there is no statistically significant difference between the mean return of the portfolios or the variances of the portfolios. Therefore, it appears that the presence of a dividend reinvestment plan has no effect on the return of a portfolio of utility stocks over the time period studied. Also, there is no statistically significant evidence that investing in a portfolio of utility stocks over the time frame examined would have produced greater than market returns per risk or equal return for less risk. CHAPTER 7 EX-DIVIDEND DAY DROP VALUES OF STOCKS USING DRPs Elton and Gruber (1970) find that the highest payout stocks exhibit a drop statistic, the ex-dividend price drop divided by the dividend, to be greater than one. Skinner and Gilster (1990) find the drop statistic is significantly greater than one for utilities but significantly less than one for nonutilities. Once utilities are separated out of the sample, high dividend yield nonutility stocks have drop statistics less than one. Skinner and Gilster suggest that some industry characteristic of utilities may be the cause of this phenomena. This study looks at one characteristic of the utility industry, its high usage of dividend reinvestment plans and seeks to determine if the value of the drop statistic stems from this trait. 7.1 Background on Stock Ex-Dividend Day Behavior When a stock goes ex-dividend, the price of the stock should fall by the amount of the dividend. However, in a world of differential tax rates between capital gains and dividends, the fall is expected to be somewhat less than the dividend to compensate for the higher tax paid on the 114 115 dividend. Campbell and Beranek (1955) find that the ex- dividend day price falls by about ninety percent of the dividend value. However, the authors also note that the price drop varies widely between individual stocks, therefore an investor can not rely upon any particular drop actually occurring. Elton and Gruber (1970) explained the ex-dividend price drop with a tax clientele hypothesis. Assuming an investor in equilibrium is indifferent between selling the stock before it goes ex-dividend at price P5, and receiving the dividend, D, while simultaneously selling the stock on the ex-dividend day at price P" the drop statistic, (P,a - P.)/D, should equal (1-tbL/(1-t3) where t, is the capital gains tax 1 rate and to is the ordinary tax rate”. As the ordinary tax bracket approaches the capital gains rate (i.e. decreases), the drop statistic approaches 1. As an investor's ordinary tax rate gets smaller, one would expect him to invest in higher payout stocks. Therefore, a positive correlation between the drop statistic and dividend yield should exist. This is the effect that Elton and Gruber find. In all but the highest dividend yielding deciles, the drop statistic can be used to estimate marginal tax brackets. Thus a tax clientele explanation for ex-day price behavior. Lewellen, Stanley, Lease and Schlarbaum (1978) find weak evidence for a tax clientele effect. Their evidence, 13See Elton and Gruber (1970 p.69) for a full derivation of the drop statistic. 116 while statistically significant, shows that as the marginal tax bracket of investor groups increases, the dividend yield of the stocks held by these groups declines. However, the decline is small relative to the tax increase. Kalay (1982) re-examines the Elton and Gruber findings by trying to correct for a potential bias in the drop statistic. Prior studies find a negative relationship- between the 8 of a stock and dividend yield, thus Kalay argues a low dividend yielding stock should have a higher expected daily price appreciation which in turn can affect the magnitude of the ex-day drop statistic. Kalay feels the relationship between a stock's B and dividend yield may be the underlying cause of the positive correlation between dividend yield and the drop statistic. When corrections for risk are made, Kalay still finds a positive relationship between dividend yield and the drop statistic. The correlation, while positive, is not statistically significant when the ex-dividend price is adjusted for average daily returns but is positive and significant when _ adjustment for market model returns is made. However, he finds no evidence of a tax induced clientele effect. Lakonishok and Vermaelen (1983) develop their own-drop statistic which relates the average price drop to the average dividend yield. They use a sample of firms in the Canadian market to test their drop statistic. They find a positive relationship between dividend yield and changes between cum and ex-dividend day prices: however, it is less 117 pronounced than the relationship found in the 0.8. Their results are inconsistent with the tax clientele hypothesis and they suggest that short term traders capturing dividends may have a significant effect on the ex-day prices. They note that tax effects probably have an impact on the ex- dividend price but there are more factors involved than a simple tax clientele argument can explain. Eades, Hess and Kim (1983) look at the ex-dividend day drop from a return standpoint. They note that if an investor pays higher taxes on dividends than capital gains, the investor will "demand a tax premium in the form of a higher pre-tax rate of return on the ex-dividend day" (p.6). They find that while this is true, the tax clientele argument is consistent only for common stocks but not preferred. Finally, Skinner and Gilster (1990) reexamine the ex- day drop by separating utilities from nonutilities. Prior studies that support the tax clientele effect found that the highest dividend yielding stocks had ex-day drops greater than one. When Skinner and Gilster split their sample into utilities and nonutilities, they find that for all dividend yields except one, utilities had drop statistics significantly greater than one and nonutilities had drop statistics significantly less than one. Furthermore, there appears to no longer be a correlation between dividend yield and the drop statistic once utilities are separated out of the sample. The authors suggest that perhaps the tax 118 clientele hypothesis is not the explanation for the ex- dividend day drops. This study will look at a sample of utility firms and separate this sample into utilities with DRPs and utilities without DRPs. The purpose is to discern whether the DRP has any bearing on the ex-dividend day behavior. 7.2 Data and Methodology The data for this section was obtained from the 1990 QBSP master files. Quarterly dividends, ex-dividend day and day before ex-dividend day prices were acquired for the years 1972 - 1989. The sample consists of utility firms that were chosen based on SIC codes from 4900 to 4939. Firms that did not have a complete set of data throughout the time frame were discarded. There were 81 firms in the final sample. The §tang§;g_§ng_zgg;§_ggtlggk was used to obtain information about which firms have DRPs. The sample was then split into firms with DRPs, firms without DRPs and for 1977-1989, firms with discount DRPs as well. The separation was done on an annual basis. Also for the years 1974 - 1976, each group was separated into quintiles based on dividend yield, with the first quintile having the highest yield. These years were chosen because each group had relatively the same number of observations. The same separation was done for 1980-1985 using the DRP and DDRP groups. 119 To calculate the drop statistic, the Elton and Gruber (1970) method is used. = 8 (16) where Pb = the cum-dividend price IQ = the ex-dividend price, and D = the dividend. The utility industry's high usage of the DRP may affect the ex-dividend price. Advantages of DRPs, such as discounts and low transaction costs, provide value to the investor. Since the investor in the DRP is receiving more than the nominal value of the cash dividend, the ex-day price may reflect this added value, thus drop by more than the value of the cash dividend. This is especially apparent if the plan has a discount feature. Even without a discount, the added value arises from the no or low transaction costs associated with the plan. Assuming the above argument is viable, the following hypothesis is proposed. Hypothesis 7-1: Ho: The drop statistic, e, is equal to 1 for all subgroups. H,: The drop statistic, e, is greater than one for the subgroup that contains DRPs and is less than one for the subgroup without DRPs. 120 To test this hypothesis, an average drop statistic for each subgroup in each year (and each quintile where applicable) is calculated using the following equation, 1 n 9.. = —— 39m (17) :5 p. where Gun equals the drop statistic for firm i at time t in group q. q equals 1,2,3 with 1 = no DRP: 2 = DRP; 3 = DDRP. A standard t-statistic is calculated in order to determine whether the drop statistic is significantly greater or less than one. 7.3 Results The average drop statistic for each subgroup in each year is reported in Table 20. As can be seen in this table, 1972 is the only year that supports the alternative of Hypothesis 7-1. For the rest of the years, all but three of the drop statistics are greater than 1 with 63% of these significant at the 5% level. Many of the nonsignificant values and those less than one are in the no DRP group. This might lead one to argue mild support for the alternative of Hypothesis 7-1. However, the sample size of approximately half of the observations is small. For this reason, the results of these observations cannot confidently 121 Table 20 Ex-Dividend Day DrOp Values Year No DRP DRP DDRP 1972 0.87443 1.08511 # -(2.42845) * (0.57198) 1973 1.08695 1.11777 (1.36591) (1.49542) 1974 1.09024 1.22530 (1.59817) (3.03620) * 1975 1.23224 1.14768 (3.84455) * (2.21533) * 1976 1.20770 1.17770 (2.69526) * (3.76306) * 1977 1.14799 1.21226 1.23192 # (1.83241) * (5.23584) * (2.66832) * 1978 1.26606 1.17923 1.31298 (3.92220) * (4.72594) * (5.63804) * 1979 1.20611 1.11273 1.36780 (1.89375) * (2.09735) * (4.11221) * 1980 1.30980 # 1.24024 1.35528 (1.54725) (3.46123) * (5.77583) * 1981 1.11339 # 1.19001 1.06202 (0.59551) (3.55282) * (1.16361) 1982 0.91337 # 1.20221 1.20916 —(o.49271) (4.26686) * (4.13799) * 1983 1.08189 # 1.25106 1.15162 (0.68862) (5.14154) * (3.83109) * 1984 1.12918 # 1.11690 1.00805 (0.80560) (2.27619) * (0.18743) t-statistics in parentheses * Significance at the 5% level # Sample size is small; n= 8—36 122 Table 20 (cont'd) Year 1985 1986 1987 1988 1989 No DRP DRP 1.35000 # 1.16112 (1.25093) (3.33761) * 1.47939 # 1.33102 (1.66357) (6.02016) * 1.93890 # 1.32340 (2.05837) * (5.19208) * 0.79324 # 1.09599 -(1.12222) (2.20272) * 0.86703 # 1.05181 -(0.44602) (1.52106) DDRP 1.19868 (4.07239) * 1.16062 (1.15794) 1.30601 (2.32509) * 1.09566 (0.81082) 1.07086 (0.42134) t-statistics in parentheses * Significance at the 5% level # Sample size is small; n= 8-36 123 be put forth as meaningful. The discount provided by some of the firms does not seem to have an effect on the drop statistic. All of the 9's are greater than one with 8 out of 12 statistically significant. Examination of the 1974—1976 subsample, where the No DRP and DRP groups are separated into quintiles, shows mixed results. These results are reported in Table 21. The highest dividend yield decile is significantly greater than one in all but one observation throughout the chosen time period. Within each year and subgroup, there appears to be no definite pattern as far as correlation between 6 and dividend yield. Most drop statistics are greater than one but are not ascending or descending strictly with the quintiles. In fact, the correlation coefficient between the drop statistic and dividend yield is .0928 in the 1976 No DRP group and 0.2873 in the 1976 DRP group with the other coefficients ranging between 0.1395 and 0.2441. The most notable trend is that the number of drop statistics that are significant declines as the dividend yield decreases. Only _ one observation, 1976 No DRP, is statistically significant in the last quintile. Similar results can be found in the 1980-1985 time period. Table 22 shows these results. Again, within years and subgroups, there is not a strong relationship between 6 and dividend yield. The correlation coefficients run between -0.0304 and 0.2503. Table 23 displays the 124 o.6.» .656. 4m 6.. .6 ooomo.u.=o.m4 mononucoumm :H mowumwumuqu ocoofi>flo umozofl: u . 6..oo.5o ...wm...u 4 .omm~..~. .omm..o.I ..mmmm.o. .omw....I ....mo.o. m¢mom.o o.o.¢.. ooamm.o .~.6o.. .oa~m.o .mmoo.. m .~m.om... .o~.~.o.I .4oomm.o. .mooom... 4 ..om4a... .mmam.o.u 0°46... mmowm.o ~a~mo.. m.m~... .m44m.. .¢.om.o 4 .~o~mm.¢. .m~m-... .~mmmm.o. 4 .mm.mo.~. 4 ..~4...~. 4 .omm.o.~. moom~.. 664.... ~m¢mo.. m...~.. ¢~m¢¢.. m4¢m~.. m .m~ooo.~. .4mmo4... 4 .oommo.~. 4 ...vmo.~. 4 .mmm..... ..m..¢.o. .~mo~.. .m.-.. omm.~.. mm..~.. ..m-.. m.omo.. ~ .omo.o.m. 4 .m.oom... 4 .m644o.4. 4 .mo~m¢.~. .mmm.~... 4 .mmmom... mooo~.. om4om.. 4oom¢.. ammom.. .oo-.. ¢aoo~.. . ...... ...... oz ...... ...... oz ...... ...... oz 6.355.. o.m. o.m. m.m. m... «.m. 4.m. mmn Dam mma oz 66...:H5o 6.6.» mo 665.6> mono .mo oooo.>.ouxm HN OHQMB 125 Ho>oa wm on» an oozmofiuflcowm4, mononucoumm :H moflumwumuqu 6.6.. 6566.5.6 666:6.z n . 6..oo.5o 4 .oommm... .mmmvm.o. 4 .om.o.~.I .~m.¢o... .mm.m~... .~mm...o. .4.~m.. m.4.... 4m.4o.o mama... .mmmm.. m.m~o.. m .m..o~.o. .mmm.o... .oomm¢.o. 4 .mo-m... 4 .omo~m.~. ..~moo.o. ~mo.o.. 465°... mm.mo.. -..... ~¢o.~.. ammo... 4 4 .~m~.4.~. ..~om4... 4 .4m.~o.~. .m..4.... 4 ....mm... 4 .~mm¢m.~. ~oo-.. .mo~... mm~.... m.o.... -amm.. o.mo~.. m .m..o4... .mm.¢o.m. .o~...... .-mmo... 4 ...mom... 4 .moooo... oo.m... .m4m~.. o.mmo.. mom.... m~m4m.. coon... ~ 4 ...mm.... .m.mmo.m. 4 .4mmo4.~. 4 ...m44.4. 4 .¢¢.¢~.o. 4 ..4460... n.4om.. ¢o~.~.. 44oom.. ..omm.. ~o4¢4.. 4~mo4.. . ozoo mzo mmoo mmo mmoo mzo 6..u:.=o ~mm. ~ma. .ma. .am. oma. omm. meQ UGM mmn oaflucflso onHM ocoofl>fia an mosam> noun woo ocmoH>Honm NN OHQMB 126 Ho>oH wm 0:» pm mocmofiuflcmflm4 momozucoumm ca mowumflumuqu 4.6.. 4:64.546 466:442 u . 6.46:.5o 4 ..mmoo.~. .mo.o.o.I .....4.o. .m.oo~.o. .o.4..o.I ..mmmm.o. o.4-.. oo.om.o ma.mo.. .~m~o.. .m~mm.o mom~... m .oommm... 4 .4mm-.~. ...m~...- .m4m.m... 4 .-nm~.~. 4 ..~4mm... moan... ~m~m~.. .mm.m.o O4mm... m~m~... 4~m-.. 4 .o.m~m... 4 .mmo.m.~. .~m....o. .4.m4m.o. .mmomm... .m.m.m.o. .4~m... o~..~.. m-.o.. 4m~4o.. .m..... mm4oo.. m 4 .4.mmo... 4 ...m.~.~. ..4mm.... .m4o4.... 4 .m.~.m.~. 4 ..m4.m.m. one»... mo~o... ammo... m.mo... mo4-.. m.mo~.. ~ 4 .o.-o.~. .mom.m... .mmm~.o.- 4 ..m44~.~. 4 .~mm4..o. 4 .o4mm4.4. ~4.o~.. ~o-... om.ma.o amoo~.. m4.-.. .44mm.. . ..on ...... 6...... ...... ..zoo ...... 6.8.5.5 mam. mma. 4mm. 4mm. mom. mam. .o.u:oo. - 6.069 127 Table 23 Correlation Coefficient Between the Drop Statistic and Dividend Yield Year No DRP DRP DDRP 1974 0.1395 0.2326 - 1975 0.1452 0.2441 - 1976 0.0928 0.2873 - 1980 - 0.2374 0.1401 1981 - 0.0469 0.1563 1982 - 0.0909 -0.0947 1983 - 0.0830 0.2503 1984 - 0.1288 0.1353 1985 - -0.0887 -0.0304 128 correlation coefficients between 9 and dividend yield for both time periods. Across years, the highest deciles have the most observations that are significant, all but two. As one moves down the deciles, the number of significant observations declines with the last decile having only three significant observations. In neither time frame studied, 1974-1976 or 1980-1986, are there any noticeable patterns over time. Each year reflects similar results to the other years in its subsample. Also, between the two subsamples, the DRP group shows no difference in behavior. One time period to note is 1984. Only one observation in the whole year in significant. 7.4 Conclusion The empirical results do not show support for the alternative of Hypothesis 7-1. Apparently the DRP plays no role in the ex-dividend day price behavior. The added value of the plan does is not the value that is reflected by the larger than one drop statistic. Perhaps some other industry characteristics are involved. CHAPTER 8 SUMMARY Dividend reinvestment plans, which allow investors the opportunity to purchase shares of the firm instead of receiving cash dividends, were first introduced in 1968. The popularity of these plans among investors has increased over time as noted in the result section of Chapter 5. Investors appear especially drawn to plans that have discounts associated with them, thus benefiting from the purchase of shares at below market price. Investor participation increased during the time of the tax deferment of reinvested dividends in 1982-1985. The tax deferment was for participants in plans of qualified utilities only. Utilities, in general, use the dividend reinvestment plans as part of dividend policy more often than nonutilities. Particularly, utilities use new issue plans which allow them a steady inflow of new capital. Participation is found to be positively related to past and present firm stock return. Interest in a firm's plan increases with the expected outlook of increasing stock return, proxied by current and past returns. Trading volume in a firm's stock is not affected by high participation levels in DRPs. The participation rates 129 130 _ are either not large enough to cause a steady shareholder base, thus few trades, or the institutional traders are using DRPs as an inexpensive way to purchase a firm's stock, especially if there is a discount. If institutional investors are buying and selling around the ex-dividend date in order to take advantage of the DRP, volume of trade may increase. However, this study shows no support for either argument. Development of a methodology that can separate the effects of individual and institutional participation in DRPs is a goal for future research. Portfolios containing only utility stocks with DRPs are unable to outperform the market or outperform similar portfolios that do not contain DRPs. Extension of this work into other industry groups may provide more substantial support for the (in)ability of the DRP to affect portfolio return. Finally, the ex-dividend day drop for utility firms is found to be greater than one. The sample is split into firms using and not using DRPs. The findings show both _groups to have drop statistics greater than one. Also, this study is consistent with the finding that the correlation between the drop statistic and dividend yield is no longer present when the utility industry is separated from the nonutility industries. The tax clientele explanation for ex-day price drops is not supported. APPENDIX 131 1‘35»: 4.: ... 88 4 2... a... 2.5: 262.952.... ... 5.1:... ......z ... «...»... ...... .....SEEE 9.62.... ...... 2.. .8. ...... 5.5... ...... ... _ 3.. VB: gonna»: .2253. ....8 ...... ...... .825... a... .6. 7.4... at...» ...... ...: 8 2.8. ...... a: 7...... a «3 .584... u 8.3.6:...— ...= 8.2.. 3...... 3.... .5... ... x Eats... ...... 2.43.13... 5...... .4 . .... .25: 9.2.6:... ...: ....— ....._.. 28.58....2 ecu—...... .. 855:3... ...... 2: ...... ...... ...... 5...... 2: 954.558 ... ......o ...... 2.... ...... .... ....» ....2: 43.8.... 8.7.8-... as. ... ......5 a 53.8.... ....m ......4... ...... ...... .... 6.44.9... ...: .2... ... ...... 5.5... a .45.. - ...... .2... ......22... age... < x_ozm._n_< .4 c .... £34.35... 2. 3....» 2: ...: _E 3.2... LIST OF REFERENCES LIST 0? REFERENCES Amihud, Yakov, and Haim Mendelson. "Liquidity and Stock Returns." Financial Analysts Journal, May—June 1986, 43-48. Asquith, Paul, and David W. Mullins, Jr. "Equity Issues and Offering Dilution.“ Journal of Financial Economics, 1986, 61-89. Baker, H. Kent, and Martha C. Johnson. "Dividend Reinvestment Plans: A Survey of Current Practices." JOurnal of the .Midwest Finance Association, 1988, 37-49. Baker, H. Kent, and Sue E. Meeks. "An Analysis of Dividend Reinvestment Plans: 1968-1988.“ Working Paper: The American university Kogod College of Business Administration, 1990. Baker, H. Kent, and William H. Seippel. “A New Look at Dividend Reinvestment Plans.“ MSU Business Topics, Summer 1980, 39-42. Baker, H. Kent, and William H. Seippel. "The Use of Dividend Reinvestment Plans by Utilities." Akron Business and Economic Review, Spring 1981, 35-41. Baker, H. Kent, and William H. Seippel. "Dividend Reinvestment Plans Win Wide Currency." Harvard Business Review, November-December 1980, 183-186. 132 133 Bar—Yosef, S. and R. Kolodny. "Dividend Policy and Capital Market Theory.“ Review of Economics and Statistics, May 1976, 181-190. . Bhattacharya, S. "Imperfect Information, Dividend Policy, and 'the Bird in the Hand' Fallacy." Bell Journal of Economics, Spring 1979, 259-270. Black, Fischer. “The Dividend Puzzle." The Journal of Portfolio Management Winter 1976, 5-8. Black, Fischer and Myron Scholes. "The Effects of Dividend Yield and Dividend Policy on Common Stock Prices and Returns." Journal of Financial Economics, May 1974, 1-22. Boyles, Jesse V., and John L. Kramer. "Individuals May Use Dividend Reinvestment Plans to Achieve Significant Tax Savings." The J0urnal of Taxation, August 1982, 114- 122. Brennan, M. J. “Taxes, Market Valuation and Corporate Financial Policy." National Tax Journal, December 1970, 417-427. Campbell, J., and William Beranek. "Stock Price Behavior on Ex-dividend Dates." Journal of Finance,‘ December 1955, 425-429. Davey, Patrick J. Diyidend_Reinyestment_£ngrams. The Conference Board, Inc., 1976. 134 Dubofsky, David A., and Leonard Bierman. "The Effect of Discount Dividend Reinvestment Plan Announcements on Equity Value.“ Akron Business and Economic Review, Summer 1988, 58-68. Eades, K. M., P. J. Hess, and E. H. Kim. "On Interpreting Security Returns during the Ex-dividend Period.“ Journal of Financial Economics, March 1984, 3-34. Elton, E. J., and M. J. Gruber. "Marginal Stockholder Tax - Rates and the Clientele Effect." Review of Economics and Statistics, February 1970, 68—74. Fetherston, Thomas A., and Kendall P. Hill. "Are DRIP Investments Special?" Journal of Business and Economic Perspectives, 1990, 29-36. Fredman, Albert J., and John R. Nichols. "New Capital Dividend Reinvestment Plans of Electric Utilities." Public Utilities Fortnightly, February 28, 1980, 19—28. Fredman, Albert J., and John R. Nichols. "Sizing Up New Capital Dividend Reinvestment Plans." California Management Review, Summer 1982, 77-84. Finnerty, John D. "New Issue Dividend Reinvestment Plans and the Cost of Equity Capital." Journal of Business Research, 1989, 127—139. Gordon, M. J. "Dividends, Earnings, and Stock Prices.“ Review of Economics and Statistics, May 1959, 99-105. Kalay, A. "The Ex-Dividend Day Behavior of Stock Prices: A Re-Examination of the Clientele Effect." Journal of Finance, September 1982, 1059-1070. 135 Hansen, Robert S., John M. Pinkerton and Arthur J. Keown. "On Dividend Reinvestment Plans: The Adoption Decision and Stockholder Wealth Effects." Review of Business and Economic Research, Spring 1985, 1—10. Harter, H. L., and D. B. Owen, eds. Selected_1ables_in Mathematical_Stati§ti§s. Markham Publishing Company, 1970. Johnson, Lindsay, and John W. Weber, Jr. "The Economic Recovery Tax Act of 1981 and Public Utilities." Public Utilities Fortnightly, March 4, 1982, 37-40. Lakonishok, J., and Theo Vermaelen. "Tax Reform and Ex— Dividend Day Behavior.“ Journal of Finance, September 1983, 1157—1179. Lewellen, W. G., K. L. Stanley, R. C. Lease, and G. G. Schlarbaum. "Some Direct Evidence on the Dividend Clientele Phenomenon." J0urnal of Finance, December 1978, 1385-1399. Lintner, J. "Distribution of Incomes of Corporations Among Dividends, Retained Earnings, and Taxes." American Economics Review, May 1956, 97—113. Litzenberger, R. H., and K. Ramaswamy. "The Effect of Personal Taxes and Dividends on Capital Asset Prices: Theory and Empirical Evidence." JOurnal of Financial Economics, June 1979, 163—196. Masulis, Ronald W. "The Effects of Capital Structure Change on Security Prices: A Study of Exchange Offers.“ The Journal of Financial Economincs, June 1980, 139—178. 136 Masulis, Ronald W., and Ashok N. Korwar. "Seasoned Equity Offerings An Empirical Investigation." JOurnal of Financial Economics, 1986, 91-118. Mikkelson, Wayne H., and M. Meagan Partch. "Valuation Effects of Security Offerings and Issuance Process.” Journal of Financial Economics, 1986, 31-60. Miller, Merton H. “Debt and Taxes." The Journal of Finance, May 1977, 261-275. I Miller, Merton H. and Franco Modigliani. "Dividend Policy, Growth, and the Valuation of Shares." Journal of Business, October 1961, 411-433. Miller, Merton H. and Kevin Rock. "Dividend Policy under Asymmetric Information.” JOurnal of Finance, September 1985, 1031-1051. Miller, Merton H. and Myron S. Scholes. "Dividends and Taxes." JOurnal of Financial Economics, December 1978, 333-364. Myers, Stewart C. "The Capital Structure Puzzle." Journal of Finance, July 1984, 575—592. Myers, Stewart C., and Nicholas S. Majluf. "Corporate Financing and Investment Decisions When Firms Have Information that Investors Do Not Have." JOurnal of Financial Economics, June 1984, 87-221. Pangano, Marco. "Trading Volume and Asset Liquidity." The Quarterly JOurnal of Economics, May 1989, 255—274. 137 Peterson, Pamela P., David R. Peterson, and Norman H. Moore. "The Adoption of New-Issue Dividend Reinvestment Plans and Shareholder Wealth.“ The Financial Review,. May 1987, 221-232. Pettway, Richard H., and R. Phil Malone. "Automatic Dividend Reinvestment Plans of Nonfinancial Corporations." Financial Management, Winter 1973, 11-18. Reilly, Raymond R., and Timothy J. Nantell. "Dividend Reinvestment." JOurnal of the Midwest Finance Association, 1979, 101-113. Richardson, Gordon, Stephan E. Sefcik, and Rex Thompson. "A Test of Dividend Irrelevance Using Volume Reactions to a Change in Dividend Policy.“ Journal of Financial Economics, 1986, 313-333. Skinner, David L., and John E. Gilster, Jr. ”Dividend Clienteles, the Tax—Clientele Hypothesis, and Utilities. The Financial Review, May 1990, 287-296. Snedecor, George. Statistical_MeLths. Iowa State College Press, 1946. Shefrin, H. M. and M. Statman. "Explaining Investor Preference for Cash Dividends.“ Journal of Financial Economics, 1984, 253-282. Watts, R. "The Information Content of Dividends." Journal of Business, 1973, 191—211. STATE UNIV LIBRQRIE “IIISIIIIIIIZIIIIIII IIIIIIII|I7I|I|I I2|I|IIIIIIII|7I|IIIII|