Dividends, Reinvestment and Bonus Shares: The Shareholders Choice.

Transcription

1 Dividends, Reinvestment and Bonus Shares: The Shareholders Choice. James S. Murray* & Michael T. Skully Monash University Abstract In this paper we investigate the relationships between ownership structure, dividend plan features and financial variables as they affect shareholder participation in dividend reinvestment and bonus share plans. Using multivariate regression we find evidence that the method used to price new share issues, financial, ownership and dividend factors all affect plan participation. We also find that ownership structure and franking credits affect shareholder choice between the types of plans. *Corresponding Author Department of Accounting & Finance Monash University Vic 3800 AUSTRALIA Ph:

2 Dividends, Reinvestment and Bonus Shares: The Shareholders Choice. Introduction Since Australian companies have offered their shareholders alternatives to receiving a cash dividend. Initially these were Bonus Share Plans (BSPs) 2 which allowed shareholders to take bonus shares in lieu of the dividend. Dividend Reinvestment Plans (DRPs), similar to those offered in the United States, were introduced to Australia in The DRP allowed shareholders to have cash dividends reinvested in the company and instead receive an issue of new shares, Skully (1982). Early adopters of these plans merely offered one type of plan but as more companies followed some began to offer both types. This paper studies shareholder participation in these schemes when companies offer both alternatives. The majority of existing DRP participation research was undertaken in the United States, which has no equivalent to the BSP. This paper s main contribution is the study of shareholder choice when offered multiple dividend alternatives. It also studies DRP and BSP participation under a full imputation system and extends our existing knowledge of the factors affecting DRP participation in Australia. Tax and Accounting Treatment Both DRPs and BSPs operate by participating shareholders electing to have cash dividends replaced by shares of an equivalent value. The plans differ in tax and accounting treatment. BSPs allow shareholders to renounce their right to the dividend and receive a bonus issue instead. With DRPs shareholders are still deemed to receive the dividend for taxation purposes as if the cash dividend was immediately applied to the purchase of new shares. For tax and accounting purposes DRPs are a distribution of profits while BSPs are a return of capital. Australia eliminated par values on shares and simplified equity accounting from 1 July 1997 which significantly reduced the accounting distinction between the two types of plan. 1 FAI Insurances Limited shareholders approved a bonus issue plan during their June 1978 Annual General Meeting, it became operational in time for their final dividend. Dare & Company Limited and Calderman Limited soon followed, also offering a bonus share alternative for their 1978 final dividends. 2 There is no standard nomenclature, companies have also offered these, and other, plans under various titles such as Dividend Selection Plan/Scheme, Share Election Plan, Bonus Election Plan and Share Investment Plan. The rules and operation of each plan in this sample determine whether each is classified as a DRP or BSP. 3 General Property Trust offered its DRP during 1981, while in 1982 CSR Limited was the first Australian company to offer dividend reinvestment. 2

3 Capital gains tax (CGT) differences between BSPs and DRPs are significant. CGT applies to assets purchased on or after 20 September BSP shares are deemed to have been acquired at the time the original shares, which generated the issue of bonus shares, were acquired. Therefore, if the original shares were acquired before 20 September 1985 then the sale of the original and BSP shares will not be subject to CGT. When the original shares were acquired on or after 20 September 1985, both those shares and the BSP shares are subject to CGT and their total cost base is the amount paid for the original shares. In contrast DRP shares are treated as purchased when the dividend is paid and the amount of dividend applied to the purchase serves as the cost base. Cash dividends and DRP shares are both taxable income while BSP shares are a return of capital and not subject to income tax. Since 1 July 1987 Australia has operated a full imputation income tax system where resident shareholders receive franking credits, attached to dividends, representing Australian company tax paid. For individual tax payers these credits can be used to offset personal tax while for corporate tax payers the credits are retained to pass on to their shareholders. As DRP participants technically still receive dividends, they also receive any credits attached. However, as BSP participants are not subject to income tax on dividends, they receive no franking credits. Dividend and Franking Credit Streaming As franking credits have different value to different classes of investors, imputation gives companies and investors an incentive to direct the credits to the shareholders able to obtain maximum value from them. This practice is known as dividend streaming and undermines the taxation principle that all benefits and costs of ownership accrue to the owner at the time when income is generated and the tax liability created (Antioch (2001)). An early streaming scheme, offered by some companies, was the dividend election plan where shareholders could choose between a fully franked dividend and an unfranked dividend of greater cash value. 4 Although companies are not required to pay dividends there is a requirement that available franking credits are attached to any dividend paid. To avoid this regulation dividend election companies simply created two classes of share. Then franked dividends were paid on one class, using up available credits, and unfranked dividends could be paid to the remaining shareholders (Antioch (2001)). 4 For example both dividends paid by Elders IXL Limited during 1988 were available either as 7.17 cents per share fully franked, or 9.5 cents per share unfranked. These dividends are not included in the sample. 3

4 Initially dividend and franking credit streaming was a by-product of BSPs. Shareholders with little use for franking credits could participate in the BSP, receiving neither dividend nor credits and allowing the company to retain credits for later distribution to other shareholders. Australian taxation law was subsequently changed to prevent dividend and franking credit streaming. The initial dividend streaming rules applied to dividends paid after 30 June 1990 under section S.160 AQCB of the Income Tax Assessment Act (ITAA 1936). This rule captures schemes which result in a smaller transfer of franking credits than would occur through the proposed dividend and require a debit to the company s franking account as if the full dividend had been paid. So an issue of BSP shares now results in a proportional reduction of the company s franking account 6 with a loss of franking credits. A weakness of the 1990 measure was that it did not address the trading in franking credits. This could be undertaken by investors without company knowledge or involvement. So a second set of anti-streaming measures explicitly identified this short term ownership issue by matching the availability of franking credits to economic ownership. 7 During 1997 a holding period was introduced rule for investors to qualify for franking credit tax rebates. Prior Research and Theory Development Relevance of Corporate Finance Theory In this paper we look to tax based dividend policy theories provide a theoretical basis for dividend plan research. 8 Tax clientele theories, Elton and Gruber (1970) and Litzenberger and Ramaswamy 5 This section was inserted by No 57 of 1990, after being announced in the 1989 Federal Budget. 6 This rule does not affect the franking accounts of BSP companies when the dividend is unfranked. Under a full imputation system company tax is only a withholding tax, once dividends are paid the total tax paid depends on the individual s marginal tax rate. Although BSP participants are not subject to personal income tax on dividends, when franking credits are lost income is still taxed at the company level. 7 The economic risks of ownership could be transferred by separating legal and economic ownership, or through the use of derivatives. The intended requirement is for investors to maintain at least 30% of the risk for 45 (90) days for ordinary (preference) shares to qualify for the credits. Assistant Treasurer s press release No AT/25, 31/12/1997. Although only introduced in the Taxation Laws Amendment Act No 4, 1999 the rule has retrospective effect for the tax year. 8 The standard dividend policy theories do not easily incorporate the operation of dividend plans. Signalling models, Bhattacharya, Sudipto, 1979, Imperfect information, dividend policy, and "the bird in the hand" fallacy, Bell journal of economics 10, and Miller, M, and K Rock, 1985, Dividend policy under asymmetric information, Journal of finance 40, , and agency models, Rozeff, Michael S., 1982, Growth, beta and agency costs as determinants of dividend payout ratios, Journal of financial research 5, and Easterbrook, Frank H., 1984, Two agency-cost explanations of dividends, American economic review 74, , are all based on management choosing the optimal dividend payout ratio. Companies offering a dividend plan ultimately have the dividend payout ratio determined by 4

5 (1980), show investors have tax based preferences for dividends or capital gains. This preference may also explain shareholder participation in BSPs over DRPs as the main difference between them is their tax treatment. DRP Participation Studies Most of the existing DRP participation research was conducted in the US, where DRPs offer a wider range of features than in Australia. One major difference is that early American DRPs used treasury shares, purchased on-market, for reinvestment. However, today US DRP companies may use new issue, treasury or combined sources. As Australian companies cannot hold treasury shares, all DRP and BSP issues are new shares. The other main difference is the extensive use of additional cash contributions in the US. Only a few Australian companies have allowed small cash contributions, often to top-up an allocation to a marketable parcel. A early survey of US companies offering new capital DRPs by Fredman and Nichols (1982) reported median participation of 8% of shares and 16% of shareholders. They did not investigate the factors affecting participation but did find that managers expected participation rates to grow. Managers believed features such as discounts and allowing partial reinvestment 9 were likely to attract greater participation. Todd and Domain (1997) studied the factors motivating shareholder participation in US DRPs. They found higher participation rates for discount DRPs and when the companies had positive returns. They observed that participation generally increased over time. This study also covered the effects of the Economic Recovery Tax Act 10 (ERTA), which affected DRP taxation for qualifying utilities ERTA resulted in a higher number of shareholders, but not a higher proportion of shares outstanding, participating in the utilities DRPs. They note this mixed result could be due to ERTA providing greater benefits to individual rather than institutional investors. Lyroudi (1999) found a significant increase in DRP participation during the ERTA period but this was not sustained, with no significant difference between the pre- and post-erta periods. Higher participation was related to institutional ownership for the market DRP and discount DRP subshareholder decisions and this may counteract the signals and agency control mechanisms dividends contain in these models. 9 Under Australian Stock Exchange listing rules, companies must offer shareholders the choice of full or partical participation. 10 ERTA exempted reinvested dividends from income tax up to the limits $750 for individuals and $1,500 for joint returns. This exemption is as close as the US has come to a BSP-type plan. 5

6 samples. Discounts had a positive effect on participation and market purchase plans showed lower rates than new issue ones. Leverage had a positive effect on participation, size had a negative effect while the company s current ratio had no effect. Australian participation rate research is limited. Wills (1983) reported early participation rates for FAI Insurances Limited s BSP averaged 84%, ranging from 65% to 97%. In contrast the initial DRP offers from Lend Lease, Leighton Holdings and CSR Limited attracted 9.7% to 33.4%. Wills (1989) looked at the discount and time period used to determine the issue price of the new shares. She found discounts positively affected participation as did the use of a later pricing period. This pricing period issue raised by Wills (1989) is unclear as she uses the terms close of books and ex-dividend interchangeably. 11 Her explanation for the later pricing period is that shareholders will prefer the issue price to be set ex-dividend. This is logical as the ex-dividend share price is typically lower so they would receive more shares. However, her classification of almost half the sample using a cum-dividend pricing period seems doubtful as very few companies actually do this. It is more likely that she meant to classify the samples before and after close of books, while this matches company practice, 12 it is hard to explain why a pricing period after the close of books would cause higher participation. An early pricing period can benefit shareholders by creating some valuable option features. As the last date to elect to participate in the plan is the close of books, shareholders can observe prices and estimate the likely issue price before deciding to participate. Dammon and Spatt (1992) identified this as an Asian option and valued it. However, it must be noted that Asian options do not provide large payoffs. Scholes and Wolfson (1989) considered the option but found it was so rarely out of the money, because of the discount, it was not worth the effort to use it To determine a fair price for the issue of DRP and BSP shares most companies issue at the weighted average share price less a discount. The average share price is usually determined over a five day period, either between the exdividend day and the close of books or immediately after the close of books. At the time of Will s paper there was a gap of five trading days between the shares going ex-dividend (before start of trade) and the close of books (after end of trade). Between January 1992 and February 1999 this period was extended to seven trading days before returning to five days. 12 The plans in this paper use pricing periods either between ex-dividend and close of books or following close of books. See Table This is an Asian option because the issue (exercise) price is an average of the share prices over the pricing period, therefore the issue price will not differ much from the prevailing market price. Adding a discount feature will increase the difference between market and issue prices so the share price would need to decrease by twice the discount for the option to move out of the money. 6

7 The potential role of the discount is highlighted by Scholes and Wolfson (1989) who describe their investment in DRPs and additional cash contributions to exploit the availability of discount shares. They document significant actual profits gained by investing in DRPs. The discount is taxed differently in the US and Australia. In the US the discount is classified as income and must be reported in tax returns. Australia, in comparison, does not tax the discount as normal income but it increases the capital gain as it lowers the cost base of DRP purchased shares. Data Our full sample of DRP and BSP issues consists of 307 dividends paid between March 1989 and September The dividends were paid by 43 companies. The limited sample size reflects our requirement that companies report each type of plan s issues separately as many dual DRP and BSP companies simply disclose the total issue resulting from both plans. Information on dividend plan share issues and the method of setting the issue price was obtained from company announcements, supplemented by company annual reports. Descriptive Statistics The descriptive statistics in Table 1 show DRPs generally attract higher participation levels than BSPs. Discounts offered range from 0-15% and most companies set the issue price by the weighted average share price over the period between ex-dividend and the close of books. Few dividend plans were underwritten by the companies in this sample. Although not reported in the table; 65 dividends were unfranked, 61 partially franked and 181 fully franked. Methodology This paper models shareholder participation in DRPs and BSPs as a decision based on company financial data and the details of the plan offered. The explanatory variables used are publicly available information when the decision is made as we want to determine what affects the decision, and not how those variables are affected by the retention of cash. We recognise shareholder heterogeneity by adding ownership variables to the analysis. Prior research has only considered one type of plan so the emphasis was on variables affecting total participation, here we consider variables which attract reinvestment in general as well as variables that affect shareholder choice between the plans. Most of the independent variables used in this paper have featured and been found significant in previous studies. Hasseldine (1988) found that an investor s participation decision could be affected 7

8 by four types of factor: these are economy wide, company, plan design, and investor factors. Like the existing research this paper concentrates on company financial factors and plan design while adding ownership structure as a proxy for investor factors. The plan features are the discount, pricing period and underwriting. Dividend features are the payout ratio and franking. Wills (1989) found larger discounts attract greater plan participation. She also found a later pricing period increased participation. Although uncommon, if management arranges underwriting for the plans it should provide a signal to the shareholders, but the nature of the signal is uncertain. 14 Franking credits reduce shareholder tax liability and therefore there is less need for liquidity to pay tax. So franking credits should be positively related to DRP participation. However, as most observations occur after the introduction of anti-streaming legislation BSP participants may prefer unfranked dividends so the credits are not lost. The two ownership variables used are number of shareholders and the percentage of shares held by the twenty largest shareholders. Both are proxies for institutional shareholdings. Dividend research often treats institutional shareholders as low tax or untaxed. However, as Dhaliwal, Erickson and Trezevant (1999) observe this proxy is far from perfect as not all low tax investors are institutions and not all institutions are low tax investors. The financial variables include the main financial ratios; price earnings ratio, book value debt to equity, market to book and six month share return. These were selected as they have been significant in prior studies. Three of these financial variables include share price, which has been a problem when used in seasoned equity offering (SEO) research due to the adverse selection problem. For example market to book is used as a proxy for growth opportunities, which should attract reinvestment. However, a high market to book ratio can also be caused by temporarily high share prices which management are trying to take advantage of in making the issue. The adverse selection issue should not affect dividend plan issues as management cannot time the issue like they do with SEOs. Therefore these variables should reflect company performance and need for capital to finance growth. 14 Underwriting could signal that the company needs to retain cash, encouraging reinvestment, or that the company has arranged an alternative so shareholders are not asked to provide the cash. 8

9 The Model We hypothesis that the reinvestment rate is a function of financial factors similar to those in previous studies as well as ownership factors and factors relevant to dividends and reinvestment in an Australian context. Accordingly we model reinvestment as; R = α + βi + ε n i = 1 Where: R = The participation ratio or relative participation ratio Participation Ratio = Dollar Value Of Dividends Reinvested Cash Dividend With No Reinvestment = Number Of Shares Issued x Issue Price Per Share Number Of Shares Outstanding x Dividend Per Share Relative Participation = BSP Participation / Total DRP & BSP Participation α = a constant β i = independent variables representing plan design features, financial and ownership factors Results Table 4 contains the models for total and DRP reinvestment. Inspection of the raw data revealed that in 86% of cases DRP exceeded BSP participation and in 75% of cases it was more than twice the BSP participation. Therefore, the total reinvestment model should be similar to the DRP model. Results for total reinvestment are similar to previous research. Significant at the 1% level are the discount, payout ratio and both ownership variables, all with a positive effect on participation, while the debt equity ratio has a negative effect. The effect of the discount is as expected and similarly significant in the other models. The ownership variables provide a curious result as Table 2 shows they are significantly negatively correlated. 15 The later tables show that high numbers of shareholders increases DRP participation while concentrated shareholding increases BSP participation. Debt s negative effect could be due to retained earnings reducing default risk and causing a wealth transfer from equity to debtholders. 15 The negative correlation is as expected. Companies with a broader shareholders base (more shareholders) are less likely to have concentrated shareholding (shares held by top twenty). 9

10 Other significant variables are the pricing period and underwriting which are significant at 5% and 10% respectively. The role of the pricing period is difficult to interpret. Although the earlier period is preferred, and this is consistent with the option approach, it is unlikely that many shareholders pay attention to the option. Underwriting has a positive effect on participation by existing shareholders. Appointing an underwriter may provide or reinforce a signal that management need to limit the cash outflow from the dividend, however, with an underwriter in place there is no need for the existing shareholders to provide reinvestment. The effect on participation could be due to existing shareholders attempting to limit a possible wealth transfer from old to new investors. The following two models show factors distinguishing between DRP and BSP participation. Model 2 provides similar results for DRP participation. The differences are that underwriting and top 20 ownership is no longer significant, while franking has a positive and significant effect. Comparing these results with BSP participation factors in model 3 reveals differences affecting shareholder choice between DRP and BSP participation. As mentioned above, companies with a large number of shareholders attract high DRP participation while companies with concentrated ownership attract high BSP participation, which suggests an institutional preference for BSPs. Franking also has different effects on DRP and BSP participation. Franking is significant and positive on DRPs while negative on BSPs. The results in model 3 are supported by model 5 which excludes dividends paid before the introduction of anti streaming laws. Combining these findings with the t-tests in table 3, which shows higher shareholder concentration in companies paying unfranked dividends, reveals dividend and dividend plan clienteles. Companies paying unfranked dividends are more likely to attract institutional investors, and these investors have a preference for BSP participation as long as the dividend is unfranked as they do not want to waste the credits under anti streaming laws. Relative BSP participation is included in models 4 and 6. The results support the differences in ownership and franking factors observed when comparing the DRP and BSP models. The additional contribution of this model is that the discount is still highly significant, showing it has a much stronger effect on BSP rather than DRP participation. Finally, table 7 presents tests for changes in participation in DRPs and BSPs following the introduction of anti streaming laws. We have found franking credits play an important role in shareholder DRP and BSP participation decisions. The anti streaming laws can reduce an investor s access to franking credits and therefore their dividend preferences. However, neither the initial introduction of these laws in 1990 nor the subsequent change in 1997 has a significant effect. As 10

11 there were delays in passing the 1997 measures into legislation a comparison is also included, again there is no significant effect. Discussion and Implications The objective of this paper was to identify how financial, ownership and plan design factors affect shareholder participation in BSPs and DRPs. Ownership concentration and plan design were found to be important while financial factors were less likely to have an effect. The choice between dividend alternatives has not been the subject of prior research. that we have found evidence of reinvestment clienteles for each type of plan shows the choice is important. Companies paying unfranked dividends tend to have a higher level of ownership concentration. These unfranked dividends are also more likely to be reinvested through BSPs than DRPs. This may be explained by the liquidity implications, with unfranked dividends the tax has not been paid by the company so it must be paid by the shareholder. As BSPs remove the income characteristics from the transaction it converts the income tax liability into a later payment of capital gains tax. Unless Australian companies have a single, homogenous, dividend clientele their dividend policy is unlikely to suit the liquidity and tax preferences of all investors. Dividend plans like BSPs and DRPs offer alternatives which allow shareholders to shape dividend policy to better fit their needs. However, with the removal of the accounting distinctions between BSPs and DRPs many companies stopped offering BSPs. The bonus share plan still offers a valuable alternative to certain tax clienteles so should still be offered, particularly by companies paying unfranked dividends. 11

12 Table 1. Descriptive Statistics Dependent Variables Min Max Median Total Participation 1.7% 96.5% 53.3% DRP Participation 0.3% 94.3% 39.4% BSP Participation 0.1% 84.9% 7.5% Relative BSP Participation 0.4% 99.6% 18.7% Independent Continuous Variables Min Max Median DDRP 0% 15% 7.5% DBSP* 0% 15% 7.5% FRANK 0% 100% 100% P_OUT SHERS** ,000 15,524 TOP % 95.9% 66.6% PE MKTBK DE RETURN -53.3% 158.3% 5.2% Independent Dummy Variable Counts 0 1 PP UND Notes: * In all but 11 dividends the same discount is offered on both plans **Here the raw numbers are presented, while the statistical analysis uses logs of these figures. DDRP & DBSP are the respective discounts offered under each plan. PP is a dummy variable for the pricing period used in determining the average share price (1 = between ex dividend and close of books, 0 = after close of books). UND is a dummy variable for the use of an underwriter to the plan (1 = underwritten, 0 = otherwise). FRANK is the franking percentage for the dividend. P_OUT is the proportion of earnings paid as a dividend before reinvestment. SHERS is the log of the number of shareholders. TOP20 is the percentage of shares held by the twenty largest shareholders. PE is the price to book ratio. MKTBK is the ratio of market to book value of equity. DE is the debt to market ratio. RETURN is the six month simple return for the share, measured at the end of the last month cum-dividend. 12

13 Table 2. Correlation Matrix for Continuous Variable DDRP DBSP FRANK P_OUT SHERS TOP20 PE MKTBK DE RETURN DDRP DBSP FRANK P_OUT SHERS TOP PE MKTBK DE RETURN Notes: Spearman correlations on upper diagonal and Pearson correlations on lower diagonal, with (2-tailed) significance levels in italics. DDRP & DBSP are the respective discounts offered under each plan. All variables are as defined in Table 1. 13

14 Table 3. Ownership t-tests Franked and Unfranked Final Dividends Shares Owned by Top 20 Shareholders Number of Shareholders Unfranked Franked Unfranked Franked Mean Mean Variance Variance Observations Observations Hypothesized Mean Difference 0 Hypothesized Mean Difference 0 df 64 df 107 t Stat t Stat P(T<=t) one-tail P(T<=t) one-tail t Critical one-tail t Critical one-tail P(T<=t) two-tail P(T<=t) two-tail t Critical two-tail t Critical two-tail Notes: Two-sample t-tests assuming unequal variances. Samples consist of fully franked and unfranked final dividends. 14

15 Table 4 Estimates of Determinates of Total & DRP Reinvestment Rates Model 1. TOTAL Reinvestment Model 2. DRP Reinvestment Coefficients t p Coefficients t p Intercept Intercept DDRP DDRP PP PP UND UND FRANK FRANK P_OUT P_OUT SHERS SHERS TOP TOP PE PE MKTBK MKTBK DE DE RETURN RETURN R R N 273 N 273 Notes: The dependent variable in Model 1 is the proportion of dividend applied to the purchase of shares through both the DRP and BSP. The DRP discount is used as dividend reinvestment is the dominant dividend plan, but as noted in Table 1 very few dividends feature different discounts on the DRP and BSP. The dependent variable in Model 2 is the proportion of dividend applied to the purchase of shares only through the DRP. 15

16 Table 5 Estimates of Determinates of BSP & Relative BSP Reinvestment Rates Model 3. BSP Reinvestment Model 4. Relative BSP Reinvestment Coefficients t p Coefficients t p Intercept Intercept DBSP DBSP PP PP UND UND FRANK FRANK P_OUT P_OUT SHERS SHERS TOP TOP PE PE MKTBK MKTBK DE DE RETURN RETURN R R N 273 N 273 Notes: The dependent variable in Model 3 is the proportion of dividend applied to the purchase of shares through the BSP. The dependent variable in Model 4 is the proportion of dividend applied to the purchase of shares through the BSP relative to the total DRP and BSP reinvestment. 16

17 Table 6 Estimates of Determinates of BSP & Relative BSP Reinvestment Rates, Post Streaming Sample Model 5. BSP Reinvestment Model 6. Relative BSP Reinvestment Coefficients t p Coefficients t p Intercept Intercept DBSP DBSP PP PP UND UND FRANK FRANK P_OUT P_OUT SHERS SHERS TOP TOP PE PE MKTBK MKTBK DE DE RETURN RETURN R R N 232 N 232 Notes: The dependent variable in Model 5 is the proportion of dividend applied to the purchase of shares through the BSP. The dependent variable in Model 6 is the proportion of dividend applied to the purchase of shares through the BSP relative to the total DRP and BSP reinvestment. The sample is limited to observations after 1-July-1990, when anti dividend streaming rules applied. 17

18 Table 7 Effects of Legislative Changes on Plan Use 1989/ /91 Anti Dividend Streaming Laws Introduced Levene's Test for t-test for Equality of Means Equality of Variances F Sig. t df Sig. (2-tailed) TOTAL DRP BSP RBSP / /98 Holding Period Rule Introduced Levene's Test for t-test for Equality of Means Equality of Variances F Sig. t df Sig. (2-tailed) TOTAL DRP BSP RBSP / /2000 Holding Period Laws Passed Levene's Test for t-test for Equality of Means Equality of Variances F Sig. t df Sig. (2-tailed) TOTAL DRP BSP RBSP Notes: t-tests of financial year to financial year changes to reinvestment rates, independent samples, assuming equal variance. 18

19 References: Antioch, Gerard, 2001, Franking credit trading, Australian tax forum 16, 323. Bhattacharya, Sudipto, 1979, Imperfect information, dividend policy, and "the bird in the hand" fallacy, Bell journal of economics 10, Dammon, Robert M., and Chester S. Spatt, 1992, An option-theoretic approach to the valuation of dividend reinvestment and voluntary purchase plans, Journal of finance 47, Dhaliwal, Dan, Merle Erickson, and Robert Trezevant, 1999, A test of the theory of tax clienteles for dividend policies, National tax journal 52, Easterbrook, Frank H., 1984, Two agency-cost explanations of dividends, American economic review 74, Elton, E, and M Gruber, 1970, Marginal stockholders' tax rates and the clientele effect, Review of economics and statistics 52, Fredman, Albert J., and John R. Nichols, 1982, Sizing up new capital dividend reinvestment plans, California management review 24, Hasseldine, D J, 1988, Dividend reinvestment schemes in New Zealand, Asian-Pacific tax and investment bulletin 6, Litzenberger, Robert H., and Krishna Ramaswamy, 1980, Dividends, short selling restrictions, tax induced investor clienteles and market equilibrium, Journal of finance 35, Lyroudi, Katerina, 1999, Significance and implications of participation in dividend reinvestment plans empirical research findings, Journal of financial management & analysis 12, Miller, M, and K Rock, 1985, Dividend policy under asymmetric information, Journal of finance 40, Rozeff, Michael S., 1982, Growth, beta and agency costs as determinants of dividend payout ratios, Journal of financial research 5, Scholes, Myron S., and Mark A. Wolfson, 1989, Decentralized investment banking: The case of discount dividend-reinvestment and stock-purchase plans, Journal of financial economics 24, Skully, Michael T., Dividend reinvestment plans: Their development and operations in Australia and the United States (Committee for the Economic Development of Australia, Melbourne). Todd, Janet M., and Dale L. Domain, 1997, Participation rates of dividend reinvestment plans: Differences between utility and nonutility firms, Review of financial economics 6, Wills, Ann, 1983, Dividend reinvestment plans. A new source of equity capital, Australian Accountant 53, Wills, Ann, 1989, A decade of dividend reinvestment, Australian accountant 59,