Dividend Imputation and Shareholder Wealth: The Case of New Zealand

Journal of Business Finance & Accounting, 29(7) & (8), Sept./Oct. 2002, 0306-686X

Dividend Imputation and
Shareholder Wealth: The Case of
New Zealand
Andrew Prevost, Ramesh P. Rao and John D. Wagster*
1. INTRODUCTION

Effective from April 1, 1988, New Zealand changed its existing two-tier `classical ' dividend taxation regime to full dividend imputation. Corporate income is now only taxed once rather than at both the corporate and shareholder level. Concurrently, the New Zealand tax code was revised, reducing the top personal and corporate income tax rates to 33 percent while standardizing the amount of tax paid on goods and services.
The New Zealand government specified several goals for the new program. A primary objective was to reduce `the influence of the tax system on firms ' financial policy and investment strategy '
(1987 Budget Speech, June 18, 1987). It was felt that without reform, it was more attractive to retain earnings than to pay dividends. Some argued this locked capital into less productive activities within existing companies, and made these companies prime targets for takeover and asset-stripping activities.
Furthermore, the old system encouraged firms to finance with
* The authors are respectively, Assistant Professor in the Department of Finance, Ohio
University; Professor of Finance in the Department of Finance, Oklahoma State
University; and Associate Professor in the Department of Finance, Wayne State University.
They wish to thank an anonymous referee whose comments led to substantial improvements in the paper; Hamish Anderson, Martin Lally, Chris Malone, Austin
Murphy, and participants at the 1998 and 1999 Southern Finance Association for their comments; and the University of Otago for generously supplying the data. Any remaining errors are the sole responsibility of the authors. (Paper received February 2001, revised and accepted August 2001)
Address for correspondence: John Wagster, 321 Aqua Court, Royal Oak, Michigan 48073,
USA.
e-mail: [e-mail] ß Blackwell Publishers Ltd. 2002, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA.

1079

1080

PREVOST, RAO AND WAGSTER

debt leading to greater indebtedness and, consequently, more highly leveraged firms subject to greater financial risk than might be socially optimal.1 To determine if investors expected New
Zealand 's full dividend imputation program to address these concerns, two questions must be addressed: (1) Did the new program encourage firms to pay dividends? and (2) Did the new program discourage debt financing?
These questions are addressed by testing two hypotheses, the imputation credit hypothesis and the tax shield hypothesis. The imputation credit hypothesis states that the present value of future imputation credits will be larger for firms with the highest expected dividend pay out and with the highest effective tax rate.
Thus, ceteris paribus, shareholder wealth gains (losses) to announcements concerning the implementation of New
Zealand 's dividend imputation program should be increasing
(decreasing) in the expected level of dividend pay out and effective tax rate. The tax shield hypothesis states that the wealth gain from the cessation of the double taxation of dividends
(which removes the tax advantage of debt) could be offset in firms that have high pre-existing levels of debt. Thus, ceteris paribus, shareholder wealth gains (losses) should be detected for low (high) debt firms to announcements concerning the implementation of dividend imputation.
The empirical evidence strongly supports the tax shield hypothesis. As displayed in Table 4, shareholders of low-debt firms experience significant wealth gains while shareholders of high-debt firms experience significant wealth losses. Moreover, these event study results are confirmed by a cross-sectional regression whose results are displayed in Panel A of Table 6. The negative sign of the highly significant debt variable reflects the distribution of wealth gains and losses displayed in Table 4. The results for the low-debt firms indicate that investors expected the elimination of the double taxation of corporate income to benefit New Zealand shareholders. The results for the high-debt firms indicate the new program also reduced the tax benefits of debt financing, which meant that these firms faced a costly reallocation of their capital structures to achieve their optimum level of debt versus equity financing in the new tax regime. This conclusion is supported by an examination of the debt ratios of
New Zealand firms. As revealed in Panel B of Table 6, the debt ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1081

levels declined between the pre-imputation and the postimputation periods. The empirical evidence is inconclusive regarding the imputation credit hypothesis.
The remainder of the paper is organized as follows: Section 2 presents the hypotheses; Section 3 describes the data and empirical methodology; Section 4 discusses the empirical results, and Section 5 concludes.

2. HYPOTHESES

(i) The Imputation Credit Hypothesis
Simply put, the purpose of full dividend imputation is to ensure that corporate profits are only taxed once. Since dividends are paid from after-tax earnings, an imputation tax credit equal to the corporate tax paid on pre-tax profits is attached to the dividend. Resident shareholders are then taxed on the `grossed up ' amount of the dividend, but they may reduce their personal tax liability with the imputation credit that reflects the full amount of corporate tax paid. Hence, shareholders pay additional tax on a fully imputed dividend only if their personal tax rate exceeds the corporate tax rate; however, these two rates were equalized at 33 percent in 1988. Shareholders in income tax brackets lower than the corporate tax bracket actually receive more imputation credits than are necessary to offset the tax liability associated with the grossed-up dividend; these additional credits may be used to offset additional non-dividend taxable income. Basically, if a New Zealand company is paying tax on its profits, it can distribute its dividend tax free to taxable New
Zealand residents.
For example, assume a firm in the 33 percent marginal tax bracket earns nine cents, which requires a three-cent tax payment. If the firm then pays a six-cent dividend, it forwards six cents in cash plus a three-cent imputation credit. If the recipient 's marginal tax bracket is also 33 percent, a nine-cent dividend will be declared for tax purposes, which produces a three-cent tax liability. However, the recipient can offset this liability with the imputation credit. Thus, for domestic tax-paying investors, the payment of New Zealand corporate tax essentially ß Blackwell Publishers Ltd 2002

1082

PREVOST, RAO AND WAGSTER

becomes a withholding tax for shareholders.
Hamson and Ziegler (1990, p. 51) argue that the new program will encourage the rearrangement of traditional dividend clienteles. Taxable New Zealand investors with marginal tax rates less than or equal to the corporate tax rate will gravitate toward high dividend pay out firms that distribute large credits. Because these shareholders prefer fully imputed dividends, they no longer have an incentive to invest in firms that minimize corporate tax, an activity that lowers the size of the imputation credit. On the other hand, investors that are unable to take advantage of the imputation credits, such as tax-exempt New
Zealand and foreign investors, should continue to prefer firms that minimize corporate tax.2 Hamson and Ziegler 's (1990) argument motivates our first hypothesis:
H1: The imputation credit hypothesis states that if taxed domestic investors are the primary price-setters in the
New Zealand market, the present value of future imputation credits will be larger for firms with the highest expected dividend pay out and the highest marginal tax rate. Thus, ceteris paribus, the market reaction to announcements pertaining to the implementation of dividend imputation should be increasing in the expected level of dividend pay out and in the firm 's marginal tax rate.
Theoretical support for this hypothesis comes from CAPMbased imputation models by Ashton (1989), Cliffe and Marsden
(1992), Lally (1992), Monkhouse (1993) and Wood (1997). They show that higher dividends imply a higher potential imputation credit (with respect to taxed domestic investors who are able to utilize the credits). Thus, the benefits from imputation are roughly proportional to the dividend payout, which suggests that the value of imputation credits will vary across firms with different payout ratios (and marginal tax rates).
For example, Ashton (1989, p. 82) shows that the benefit of owning shares in a dividend imputation system should be represented by an expression for the cost of capital which specifically accounts for the average value of useable imputation credits to shareholders. He uses a mean-variance pricing model to develop a modified form of the CAPM that illustrates the ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1083

impact of the UK advanced corporate tax (A.C.T.) imputation tax system on expected returns:
E…rj † ˆ rf …1 À T † ‡ j ‰E…rm † À rf …1 À T †Š

…1†

where E…rj † is the required return on asset j, rf is the return on the risk-free asset, T is the effective corporate tax rate, rm is the market return, and j is asset j 's market beta. However, as Ashton points out, this model is highly dependent on the assumptions made regarding the tax implications of corporate debt. For example, Cliffe and Marsden (1992) arrive at a similar afterpersonal tax equilibrium-pricing model by assuming equivalent personal tax rates on debt and dividend income and invoking the conditions of dividend irrelevancy. In such a case, it is clear that the effect of imputation is to lower the cost of capital and increase after-tax equity returns.
(ii) The Tax Shield Hypothesis
Officer 's (1994) model of the implications to a firm 's cost of capital from the Australian program of full dividend imputation suggests that shareholders of some firms may actually experience a wealth loss because of dividend imputation. This would occur if the reduction in the value of the debt tax shield outweighs the increase in firm value resulting from the elimination of the double taxation of corporate income.
Officer (1994) posits that adjusting net cash flows can capture the impact from imputation and/or the WACC for the value of the imputation credits. In Officer 's (1994) model, credits are valued relative to the extent that they may be utilized (i.e. used to reduce personal tax liability) by the investor. In particular, the relevant definition of after-tax cash flows in an imputation system involves adding back the value of the imputation credit as well as removing the tax deductibility of debt. Officer (1994, p. 7) shows that the WACC can be defined in the following way:
S
D
‡ rD …1 À T …1 À †† X
…2†
V
V
The first term reflects the cost of equity and the second term reflects the after-tax cost of debt. In an imputation system where full (100 percent) credit is given for corporate tax paid (i.e. r ˆ rE

ß Blackwell Publishers Ltd 2002

1084

PREVOST, RAO AND WAGSTER

where  ˆ 1), it is clear that debt is less effective as a tax shield and is completely neutral with respect to equity if 100 percent of corporate profits are credited to shareholders.3 Officer 's (1994) argument motivates our second hypothesis:
H2: The tax shield hypothesis states that if taxed domestic investors are the primary price-setters in the New
Zealand market, the value gained from the move to imputation may be more than offset in firms that have high pre-existing levels of debt because full imputation removes the tax advantage from using debt. Thus, ceteris paribus, the market reaction to announcements pertaining to the implementation of dividend imputation should be decreasing in the proportion of debt in the capital structure.
In summary, assuming that taxed domestic shareholders are able to set equity prices, our hypotheses suggest that the value of imputation is clearly linked to the extent to which a firm pays corporate tax, the proportion of its earnings paid out as fully imputed cash dividends and its debt level. Consequently, the introduction of a full imputation tax system should, ceteris paribus, benefit shareholders of firms with low debt levels that provide large imputation credits. On the other hand, because the loss of interest tax shields may offset the value of imputation credits for some firms; the new program may disadvantage some shareholders, particularly shareholders of firms with high debt levels that provide small imputation credits.

3. DATA AND EMPIRICAL METHODOLOGY

(i) Data
To the authors ' knowledge, this is one of the first empirical studies of the consequences of implementing dividend imputation in New Zealand. This is probably because dividend imputation was introduced prior to 1990, when returns data on many New Zealand companies were not readily available. For this study, pre-1990 data were obtained from the NZ Share Price
Database maintained by the University of Otago, which provides ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1085

daily closing prices, number of days since a stock 's last trade, dividend information, and other variables. An equal-weighted daily market index was constructed from the returns of all listed firms to adjust returns for market risk. Historical accounting data for the study were obtained from Datex Investor Services.4
During the 1980s, the New Zealand market was highly volatile and was characterized by a significant number of companies being listed and de-listed from one year to the next. To control for any problems arising from this characteristic, only firms that were continuously traded between 1983 and 1988 are used.
Moreover, the NZSE is small and many stocks are infrequently traded; we maintain a minimal level of liquidity by imposing a restriction that a stock has to trade at least once over all five event days analyzed. These two restrictions result in an initial sample size of 92 firms. While this sample selection process does not eliminate all possible bias caused by low liquidity, it does result in a sample of primarily large firms that should be reflective of the rest of the market.
A concern in conducting an event study for firms that trade infrequently in a small market is sample size. All one can do is use all available firms and hope that the sample size will prove adequate for every test. In this study, three tests are conducted that require our 92 firm sample to be divided into four subsamples per test. This results in one portfolio of 14 firms, five portfolios of 20 firms, and three portfolios of 21 and 26 firms. By way of comparison, Wagster (1996), who also conducts a regulatory event study, uses seven portfolios of banks from various countries. These portfolios contain 22, 10, 7, 6, 6, 5 and
3 firms, for a total of 59. Therefore, there is no reason to suspect that the empirical work has been unduly compromised because of the number of firms available for study during this time period.
To test the imputation credit hypothesis, the 92 firms are sorted into four portfolios based on five-year average dividend yields
(dividend/stock price; each annual yield is the average of the semi-annual yields for that year) for 1984 through 1988; and by average effective tax rate (AETR) for 1988 ((before tax profit À net profit)/before tax profit).5 To test the tax shield hypothesis, the sample is sorted into four portfolios based on the 1988 debt ratio ((long-term debt + current liabilities)/total assets).

ß Blackwell Publishers Ltd 2002

1086

PREVOST, RAO AND WAGSTER

(ii) Methodology
Many studies of regulatory events exclusively use macroeconomic and accounting data. Schwert (1981), however, argues stockmarket data measures the impact of regulatory change better than other measures because asset prices incorporate all relevant information as soon as it becomes available. Hence, we follow the large and growing literature that studies the impact of changes in regulation by measuring changes in shareholder expectations.
Works by James (1983), Allen and Wilhelm (1988), Eyssell and
Arshadi (1990), Cooper, Kolari and Wagster (1991), and Wagster
(1996) are a representative sample of this literature.
A potential problem when studying regulatory events with stock market data is that the firm returns may be contemporaneously correlated because they are exposed to the same economic event.
Therefore, we follow Schwert (1981), Dann and James (1982) and Cooper, Kolari and Wagster (1991) who observe that potential event date clustering problems can be mitigated by using equal-weighted portfolios of stock returns.
A comprehensive search for imputation articles in the two most widely read newspapers in New Zealand, the New Zealand Herald and the National Business Review, was conducted. Because most news articles containing information pertaining to the new imputation program refer to several New Zealand government press releases and consultative documents as their source, we decided these should be used as events. The probability of unanticipated news being released to the market by the original government announcement seems much higher than for a newspaper report about the government announcement that was only available at a later date. Moreover, this methodology is consistent with much of the regulatory event study literature, such as Eyssell and Arshadi (1990) who study the impact of
Federal Reserve issued press releases on US bank stock returns.
Both three-day and seven-day event windows are used because of the diffuse nature of government announcements. Additionally, both papers were perused, because neither was indexed over the study 's time period, for any possible news stories about regulations or firm-specific reports that might contaminate the event windows. No contaminating announcements were found.
The Appendix provides a list of the five events and their dates. ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1087

Both three-day …À1Y ‡1† and seven-day …À1Y ‡5† event windows are used to measure wealth changes to regulatory announcements. Abnormal returns around each event are measured by calculating mean daily returns for each portfolio and then using standard event study methodology to identify abnormal portfolio returns. We use a single index market model
(SIMM):
Rit ˆ i ‡ i RMt ‡ 4it Y where: Rit
RMt
i
i
4it

ˆ
ˆ
ˆ
ˆ
ˆ

…3†

stock return of portfolio i in time t; return of the market index in time t; portfolio i 's constant component of return; sensitivity of portfolio i to the market index; nonsystematic component of portfolio i 's return.

Mean returns are computed for the entire five-year time period for each portfolio. SIMM estimation period parameters are obtained by removing …À10Y ‡10† windows surrounding each event and appending them to the end of each of the portfolios ' time series of returns. Thus, abnormal returns for each portfolio are based on estimation period parameters from the time series of average returns, excluding the 21-day window surrounding each event. The result of this process is a 958-day estimation period, followed by five 21-day event windows for which daily abnormal returns are estimated for each of the four portfolios:
” ‡ ”i RMt †X
ARi ˆ Rit À …

…4†

Cumulative abnormal returns (CARs) for the …À1Y ‡1† and
…À1Y ‡5† windows around each event date are then calculated from daily abnormal returns generated from the 958-day estimation period.

4. EMPIRICAL RESULTS

(i) Event Study Results
Table 1 presents the impact of each of the five events on the complete 92-firm sample, as well as the cumulative impact of all of the events. Three-day excess returns are presented in Panel A, ß Blackwell Publishers Ltd 2002

1088

PREVOST, RAO AND WAGSTER

Table 1
Shareholder Wealth Effects to Events Regarding the Impending
Change to Full Dividend Imputation in New Zealand
Event

CAR (%)

t-statistic

À0.65
À0.13
0.48
0.83
0.45
0.98

À0.51
À0.11
0.38
0.64
0.35
0.34

À0.04
À0.36
0.84
À1.59
À0.89
À2.05

À0.02
À0.18
0.43
À0.80
À0.45
À0.46

Panel A: …À1Y ‡1† Window
1. August 20, 1985
2. April 13, 1987
3. June 18, 1987
4. December 17, 1987
5. April 1, 1988
Cumulative CAR
Panel B: …À1Y ‡5† Window
1. August 20, 1985
2. April 13, 1987
3. June 18, 1987
4. December 17, 1987
5. April 1, 1988
Cumulative CAR

Notes:
Panels A and B present cumulative abnormal returns (CARs) and t-statistics for a portfolio of 92 New Zealand firms for each of five events. The CARs in Panel A (Panel B) are based on three-day (seven-day) event windows surrounding each announcement. The cumulative wealth effect corresponding to all five announcements is also indicated for each portfolio. * Denotes significance at the 10% level, ** denotes significance at the 5% level, and *** denotes significance at the 1% level.

and seven-day excess returns are presented in Panel B. From the table, it is apparent that the overall sample does not exhibit any discernable reaction to any of the events. Panels A and B also indicate that the cumulative impact over all five events is not significantly different from zero for both event windows.
Tables 2 and 3 provide direct tests of the imputation credit hypothesis, while Table 4 provides a direct test of the tax shield hypothesis. Specifically, Table 2 reports the significance of the announcement effects corresponding to three- and seven-day event windows for four portfolios sorted by five-year average dividend yield, Table 3 for portfolios sorted by AETR, and Table
4 for portfolios sorted by debt ratio. Essentially, our hypotheses stipulate opposite event period wealth effects for the results of the portfolios presented in Tables 2 and 3 from those presented in Table 4. The imputation credit hypothesis stipulates that the ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1089

Table 2
Shareholder Wealth Effects to Events Regarding the Impending
Change to Full Dividend Imputation: Five-Year Average Dividend Yield
Sorted Portfolios
Column 1
…À1Y ‡1† Window
Event
Portfolio 1
(zero dividend)

Cumulative CAR
Portfolio 2

Cumulative CAR
Portfolio 3

Cumulative CAR
Portfolio 4
(highest dividend)

Cumulative CAR

Column 2
…À1Y ‡5† Window

CAR (%)

t-statistic

CAR (%)

t-statistic

1
2
3
4
5

À2.44
0.53
1.92
0.39
3.22
3.62

À0.53
0.12
0.41
0.08
0.69
0.35

1.13
1.46
2.13
À7.87
5.97
2.82

0.16
0.21
0.30
À1.11
0.84
0.18

1
2
3
4
5

0.19
À0.99
1.32
À1.08
À1.11
À1.67

0.12
À0.64
0.84
À0.69
À0.71
À0.48

À0.24
À1.66
2.31
1.02
À3.50
À2.07

À0.10
À0.70
0.97
0.43
À1.47*
À0.39

1
2
3
4
5

0.45
0.03
À0.61
À0.45
0.54
À0.04

0.37
0.03
À0.50
À0.37
0.45
À0.01

0.28
À0.36
0.31
À6.97
0.37
À6.37

0.16
À0.20
0.17
À3.74***
0.20
À1.53*

1
2
3
4
5

À1.30
0.26
À0.27
3.83
0.19
2.71

À0.59
0.12
À0.13
1.73**
0.09
0.55

À0.78
0.14
À1.23
5.39
À2.78
0.74

À0.23
0.04
À0.36
1.59*
À0.82
0.10

Notes:
The five-year average dividend yield (dividend / stock price; each annual yield is the average of the semi-annual yields for that year) is from 1984 through 1988. The first
(second) column presents three-day (seven-day) CARs (in percent) and t-statistics for four portfolios of low- to high-average dividend yield firms for each of five events. The CARs are based on three-day (column 1) and seven-day (column 2) event windows surrounding each announcement. The cumulative wealth effect corresponding to all five announcements is also indicated for each portfolio. Portfolio 1 consists of 14 firms,
Portfolios 2±4 consist of 26 firms. * Denotes significance at the 10% level, ** denotes significance at the 5% level, and *** denotes significance at the 1% level.

ß Blackwell Publishers Ltd 2002

1090

PREVOST, RAO AND WAGSTER

impact of imputation should be monotonically increasing in dividend yield and average effective tax rate (AETR), so that the most positive wealth impact should be associated with the higher dividend yield and higher tax paying portfolios, and vice versa.
On the other hand, the tax shield hypothesis stipulates that the wealth impact of imputation should be monotonically decreasing in percentage of debt in the capital structure, so that the most positive wealth impact from imputation should be associated with the lowest debt portfolio, and vice versa.
Turning to the imputation credit hypothesis, column 1 (2) of
Table 2 illustrates the three-day (seven-day) wealth effect for each of the five events. Portfolio 4 has a significant shareholder wealth increase at the 5 percent level to event 4 for the three-day event window, and a significant wealth increase at the 10 percent level for the seven-day window. Even though the cumulative CARs also have the hypothesized correct positive sign, neither is significant.
Portfolio 3, for the seven-day window, experiences a significant wealth loss at the 1 percent level to event 4 and also experiences a significant cumulative wealth loss at the 10 percent level.
Portfolio 2 experiences a significant wealth loss at the 10 percent level to event 5 for the seven-day window. The negative sign of the cumulative CARs for Portfolios 2 and 3 support the hypothesis because wealth losses are predicted as the dividend yield of the portfolios ' decrease. However, there is only one weakly significant cumulative CAR detected for any of the portfolios, and no discernible wealth effect is detected for Portfolio 1, whose results for the most part also have the incorrect sign. Therefore, while there is some evidence in support of the hypothesis, it is inconclusive and not consistent across all four portfolios.
Table 3 presents results for portfolios sorted by AETR. The fourth announcement produces highly significant shareholder wealth losses for Portfolios 1, 2, and 3 using the seven-day event window, which, especially for the first two portfolios, provides support for the imputation credit hypothesis. However, the fifth event produces a significant wealth gain for Portfolio 3 in both event windows, offsetting the wealth effect detected for event 4.
Moreover, no significant wealth effects are detected for Portfolio
4, and none of the portfolios have a significant cumulative CAR.
Therefore, as in the case of the dividend yield-sorted portfolios, while there is some evidence in support of the imputation credit ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1091

Table 3
Shareholder Wealth Effects to Events Regarding the Impending
Change to Full Dividend Imputation: Average Effective Tax Rate
(AETR) Sorted Portfolios
Column 1
…À1Y ‡1† Window
Event
Portfolio 1
(zero AETR)

Cumulative CAR
Portfolio 2

Cumulative CAR
Portfolio 3

Cumulative CAR
Portfolio 4
(highest AETR)

Cumulative CAR

Column 2
…À1Y ‡5† Window

CAR (%)

t-statistic

CAR (%)

t-statistic

1
2
3
4
5

À0.30
0.05
1.70
À1.15
À0.80
À0.50

À0.21
0.04
1.14
À0.78
À0.54
À0.15

1.60
0.23
0.56
À6.21
1.92
À1.90

0.70
0.10
0.25
À2.73***
0.85
À0.37

1
2
3
4
5

0.63
À0.60
À0.62
À0.73
1.37
0.05

0.30
À0.28
À0.29
À0.34
0.64
0.01

1.79
À2.65
0.43
À6.28
1.16
À5.55

0.55
À0.81
0.13
À1.92**
0.35
À0.76

1
2
3
4
5

À0.57
0.06
À0.42
À0.11
2.15
1.10

À0.49
0.06
À0.36
À0.10
1.83**
0.42

À0.04
À0.28
1.54
À5.22
4.81
0.81

À0.13
À0.16
0.86
À2.90***
2.67***
0.20

1
2
3
4
5

À0.67
À1.67
0.08
1.00
À1.06
À2.32

À0.37
À0.90
0.05
0.55
À0.58
À0.56

0.10
À1.16
0.19
0.16
0.08
À0.63

0.04
À0.41
0.07
0.06
0.03
À0.09

Notes:
AETR is the average effective corporate tax rate for 1988 ((before tax profit À net profit)
/ before tax profit). The first (second) column presents three-day (seven-day) CARs (in percent) and t-statistics for four portfolios of low- to high-average AETR firms for each of five events. The CARs are based on three-day (column 1) and seven-day (column 2) event windows surrounding each announcement. The cumulative wealth effect corresponding to all five announcements is also indicated for each portfolio. Portfolios 1±3 consist of 20 firms, and Portfolio 4 consists of 21 firms. * Denotes significance at the 10% level, ** denotes significance at the 5% level, and *** denotes significance at the 1% level.

ß Blackwell Publishers Ltd 2002

1092

PREVOST, RAO AND WAGSTER

hypothesis, it is inconclusive and not consistent across all four portfolios. Table 4 presents results for portfolios sorted by debt ratio in order to test the tax shield hypothesis. This hypothesis asserts that because full dividend imputation removes the tax advantage of debt, the wealth gain from the cessation of the double taxation of dividends could be offset in firms that have high pre-existing levels of debt. To support this hypothesis, shareholder wealth increases (decreases) should be detected for low (high) debt firms to announcements concerning the implementation of dividend imputation. These results are indeed what we observe for both the three-day and seven-day event windows in Table 4.
With respect to Portfolio 1 (the lowest debt portfolio), most of the wealth effects have the correct positive sign, with the third and fifth event having the greatest impact. The third event causes a significant wealth increase at the 5 percent level of 4.32 percent for the three-day window, while the fifth event causes a 3 percent
(4.57 percent) shareholder wealth gain in the three-day (sevenday) event windows, both significant at the 10 percent level. The three-day (seven-day) cumulative CAR over all five events is 10.03 percent (10.16 percent), which is significant at the 5 percent (10 percent) level.
Most of the wealth effects for Portfolio 2 also have the correct positive sign. The second event causes a significant wealth gain at the 10 percent level of 1.87 percent (2.90 percent) for the threeday (seven-day) window; and the third event causes a significant wealth gain at the 10 percent level for the seven-day window.
Overall, the cumulative CAR of all five events is 1.72 percent
(7.44 percent) for the three-day (seven-day) window, although only the latter window is significantly different from zero at the
10 percent level.
In contrast, the cumulative CARs of Portfolio 3 have the correct negative sign for both the three-day and seven-day windows; however, neither is statistically significant. The fourth event, however, produces a significant wealth loss of À2X08 percent
(À2X70 percent) for the three-day (seven-day) window at the 5 percent (10 percent) level.
Finally, the results for Portfolio 4 (the highest debt portfolio) all have the correct negative sign except for the first announcement. Moreover, this portfolio has a statistically ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1093

Table 4
Shareholder Wealth Effects to Events Regarding the Impending
Change to Full Dividend Imputation: Debt Ratio Sorted Portfolios
Column 1
…À1Y ‡1† Window
Event
Portfolio 1
(lowest debt)

Cumulative CAR
Portfolio 2

Cumulative CAR
Portfolio 3

Cumulative CAR
Portfolio 4
(highest debt)

Cumulative CAR

Column 2
…À1Y ‡5† Window

CAR (%)

t-statistic

CAR (%)

t-statistic

1
2
3
4
5

À0.71
2.06
4.32
1.36
3.00
10.03

À0.35
1.00
2.09**
0.66
1.45*
2.17**

1.34
1.79
2.52
À0.06
4.57
10.16

0.43
0.57
0.80
À0.02
1.45*
1.44*

1
2
3
4
5

À0.75
1.87
0.15
À0.39
0.84
1.72

À0.57
1.41*
0.11
À0.30
0.64
0.58

1.76
2.90
2.70
À2.23
2.31
7.44

0.87
1.43*
1.33*
À1.10
1.14
1.64*

1
2
3
4
5

0.51
0.20
À0.90
À2.08
1.19
À1.08

0.47
0.18
À0.82
À1.90**
1.08
À0.44

1.29
0.36
À0.87
À2.70
0.60
À1.32

0.77
0.22
À0.52
À1.61*
0.36
À0.34

1
2
3
4
5

0.15
À2.22
À2.53
À0.87
À2.18
À7.65

0.08
À1.23
À1.40*
À0.48
À1.21
À1.90**

0.62
À2.18
À2.13
À1.44
À5.58
À10.71

0.23
À0.79
À0.77
À1.44
À2.02**
À1.74**

Notes:
The debt ratio is for 1988 and is defined as ((long-term debt + current liabilities)/ total assets). The first (second) column presents three-day (seven-day) CARs (in percent) and tstatistics for four portfolios to each of five announcements, ranked from low to high debt ratios. The CARs are based on three-day (column 1) and seven-day (column 2) event windows surrounding each announcement. The cumulative wealth effect corresponding to all five announcements is also indicated for each portfolio. Portfolio 1 consists of 21 firms, Portfolios 2±3 consist of 20 firms, and Portfolio 4 consists of 21 firms. * Denotes significance at the 10% level, ** denotes significance at the 5% level, and *** denotes significance at the 1% level.

ß Blackwell Publishers Ltd 2002

1094

PREVOST, RAO AND WAGSTER

Table 5
Mean (Median) Values of Variables
Panel A: Portfolios Formed on Five-Year Average Dividend Yield
Variable

Portfolio 1
(zero div.)

Portfolio 2

Portfolio 3

Portfolio 4
(high div.)

Debt

0.311
(0.100)

0.540
(0.546)

0.574
(0.594)

0.472
(0.438)

5-year average dividend yield

0.000
(0.000)

0.022
(0.018)

0.026
(0.026)

0.047
(0.045)

AETR

0.156
(0.000)

0.301
(0.266)

0.285
(0.293)

0.106
(0.358)

LSIZE

8.720
(9.248)

11.176
(11.368)

10.527
(10.074)

9.708
(10.004)

LBM

6.636
(7.325)

9.172
(9.190)

8.674
(8.547)

7.700
(7.865)

12

20

21

25

Number of firms

Panel B: Portfolios Formed on Average Effective Tax Rate (AETR)
Variable

Portfolio 1
(lowest tax)

Portfolio 2

Portfolio 3

Portfolio 4
(highest tax)

Debt

0.448
(0.519)

0.460
(0.477)

0.522
(0.505)

0.531
(0.571)

5-year average dividend yield

0.026
(0.021)

0.022
(0.014)

0.031
(0.029)

0.030
(0.022)

AETR

À0.884
(À0.102)

0.032
(0.000)

0.320
(0.317)

1.121
(0.534)

LSIZE

9.944
(9.695)

9.459
(9.920)

10.995
(10.555)

10.157
(9.864)

LBM

8.488
(8.472)

7.754
(7.934)

8.521
(8.318)

7.962
(8.068)

18

19

20

21

Number of firms

ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1095

Table 5 (Continued)
Panel C: Portfolios Formed on Debt Ratios
Variable

Portfolio 1
(low debt)

Portfolio 2

Portfolio 3

Portfolio 4
(high debt)

Debt

0.152
(0.182)

0.413
(0.412)

0.601
(0.601)

0.800
(0.772)

5-year average dividend yield

0.022
(0.010)

0.029
(0.022)

0.028
(0.025)

0.032
(0.027)

AETR

0.166
(0.000)

0.331
(0.353)

À0.064
(0.270)

0.422
(0.053)

LSIZE

10.063
(9.981)

10.936
(10.977)

10.632
(10.243)

8.914
(8.989)

LBM

7.793
(7.931)

8.692
(8.668)

8.686
(8.665)

7.478
(7.438)

19

20

20

19

Number of firms

Notes:
Complete accounting data were not available for a number of firms that are present in the event study tests. In Panel A, two firms are omitted from Portfolios 1 and 4 because of incomplete records. In Panel B, two firms are omitted from Portfolio 1; six firms are omitted from Portfolio 2; five from Portfolio 3; and one from Portfolio 4. In Panel C, two firms are omitted from Portfolio 1 and one firm from Portfolio 2. Debt is for 1988 and is defined as ((long term debt + current liabilities)/total assets). The five-year average dividend yield (dividend / stock price; each annual yield is the average of the semi-annual yields for that year) is for 1984 through 1988. AETR is the 1988 average effective tax rate and is approximated as ((before tax profit À net profit)/before tax profit). LSIZE is the log of the market value of equity for 1988, and LBM is the log of the book-to-market value of equity for 1988.

significant wealth loss at the 10 percent level of À2X53 percent to event 3 for the three-day window, and a statistically significant wealth loss at the 5 percent level of À5X58 percent to event 5 for the seven-day window. More importantly, Portfolio 4 's cumulative
CAR is À7X65 percent for the three-day event window, and À10X71 percent for the seven-day event window, both statistically significant at the 5 percent level.
In summary, the results displayed in Table 4 provide strong support for the tax shield hypothesis. Shareholders of firms in the lowest debt portfolio experience statistically significant cumulative wealth gains, while shareholders of firms in the highest debt portfolio experience statistically significant ß Blackwell Publishers Ltd 2002

1096

PREVOST, RAO AND WAGSTER

cumulative wealth losses. Moreover, the signs of the cumulative
CARs for Portfolio 2 and 3 are what we expected, and each portfolio experiences significant announcements, all with the correct signs.
(ii) Portfolio Characteristics
Table 5 provides further insight into the event study results revealed in Tables 2, 3 and 4. Panel A displays descriptive statistics for the four dividend yield-sorted portfolios, Panel B for the four AETR-sorted portfolios, and Panel C for the four debtsorted portfolios. In addition to these three variables, we also include two control variables that are commonly associated with firm performance in the US and elsewhere (e.g., Fama and
French, 1992; and Chen and Zhang, 1998). LSIZE is the log of the 1988 market value of equity, while LBM is the log of the 1988 book-to-market value of equity.
Turning to Panel A, firms sorted by dividend yield do not exhibit any noticeable trends in debt ratio, AETR, LSIZE or LBM across the four portfolios. In Panel B, firms sorted by AETR do not exhibit any noticeable trends except in the case of the debt ratio. Debt ratio increases within a very small range, while AETR increases within a much larger range. Since the variability of
AETR is so much greater than for the debt ratio, the significance of this trend is questionable. The negative values of AETR reflect firms that had built up tax credits from previous years so that before tax profit is less than net profit. In Panel C, firms sorted by debt ratio show no discernible trends with the other variables except a slight positive correlation with dividend yield. The significance of this trend appears inconsequential since the variability of the debt ratio is so much greater than for the dividend yield. In summary, the results displayed in Table 5 provide insight into the levels of the variables we study, and indicate that none of the variables exhibit undue dependencies.
(iii) Cross-sectional Regression Results
A cross-sectional regression analysis is conducted to obtain additional insight into the relationship between the explanatory variables and the wealth effects associated with the advent of full ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1097

dividend imputation.6 The following regression model is estimated: three-day (seven-day) CAR ˆ f (dividend yield, debt, average effective tax rate, size, book-to-market ratio).

(5)

Equation (5) is estimated using ordinary least squares (OLS) methodology where the dependent variables are …À1Y ‡1† and
…À1Y ‡5† window SIMM abnormal returns for each firm corresponding to each event. The five-year average dividend yield (dividend/stock price; each annual yield is the average of the semi-annual yields for that year) is for 1984 through 1988.
The debt ratio (DEBT) is for 1988 and is defined as ((long term debt + current liabilities)/total assets). The average effective tax rate (AETR) is for 1988 and approximated as ((before tax profit
À net profit)/before tax profit). LSIZE is the log of the market value of equity for 1988, and LBM is the log of the book-tomarket value of equity for 1988.
Panel A of Table 6 presents the cross-sectional regression results, which shows that the only significant variable in either regression is the debt ratio. It has a large negative coefficient in both models and is highly significant at the 1 percent level.7 This provides strong support for the tax shield hypothesis, which suggests that the loss of interest tax shields may offset any value gained from dividend imputation for high-debt firms. The negative coefficient for DEBT displayed in Panel A confirms the relationship demonstrated in Table 4. Recall that the event study results when applied to four debt-sorted portfolios revealed that shareholders of low-debt firms experienced significant wealth gains while shareholders of high-debt firms experienced significant wealth losses.
(iv) Debt Ratio Difference in Means (Medians) Tests of Pre- and Postimputation Periods
While the results of Table 4 and Panel A of Table 6 reflect investor expectations at the time New Zealand 's dividend imputation program was implemented, these tests provide no evidence regarding the impact of the regulation on an ex-post ß Blackwell Publishers Ltd 2002

1098

PREVOST, RAO AND WAGSTER

Table 6
Cross Sectional Regression and Debt Ratio Comparison
Intercept

DEBT

DIVYIELD

AETR

LSIZE

LBM

Panel A: Cross Sectional Regression of Abnormal Returns on Explanatory Variables
Three-day
CAR Model

À0.011
(À0.27)

À0.084
(À3.26)***

0.120
(0.35)

0.007
(1.59)

0.005
(1.38)

0.006
(0.99)

Seven-day
CAR Model

À0.001
(À0.05)

À0.050
(À2.59)***

0.091
(0.35)

0.005
(1.56)

0.002
(0.85)

À0.003
(À0.61)

Panel B: Debt Ratio Difference in Means (Medians) Tests of Pre- and Post-Imputation
Periods
1. All Available
1985±1988
0.49
(0.51)
N ˆ 79

Firms with Data Available for that Year
1991
1992
1993
1994
0.40***
0.42**
0.41**
0.41**
(0.43)*** (0.41)**
(0.45)**
(0.40)**
N ˆ 74
N ˆ 76
N ˆ 80
N ˆ 86

2. Firms with Data Available from 1985 through 1995
0.43
0.41
0.49
0.41
0.37
(0.40)
(0.45)
(0.40)
(0.38)
(0.35)
N ˆ 24
N ˆ 24
N ˆ 24
N ˆ 24
N ˆ 24

1995
1991±1995
0.41**
0.41***
(0.43)**
(0.43)***
N ˆ 92
N ˆ 408
0.37
(0.37)
N ˆ 24

0.41
(0.38)
N ˆ 120

Notes:
In Panel A, results are displayed from cross sectional OLS regressions of the three-day and seven-day cumulative abnormal returns (CARs) surrounding each event on independent variables for dividend yield, debt ratio, AETR, and size and book-to market equity control variables. For each model, coefficients are shown first and the associated t-statistic is then shown in parentheses. The F-value (R-square) for the three-day model is 3.10 (0.18), and for the seven-day model is 2.40 (0.14). In Panel B, difference-in-means tests are conducted against the 1985±1988 average debt ratio. In some cases, the 1985 debt ratio was unavailable. The first test uses all available firms for which data is available for that year.
Hence, the largest firms traded for that year on the New Zealand stock market are represented, but the same firms are not necessarily represented each year. Because the total number of ordinary shares listed on the NZSE for 1988 is 234, 1991 is 111, 1992 is
110, 1993 is 125, 1994 is 137, and 1995 is 132, the responses represent 34 percent of all firms listed on the NZSE in 1988, 67 percent in 1991, 69 percent in 1992, 64 percent in
1993, 63 percent in 1994, and 70 percent in 1995. The second test uses the 24 sample firms for which data could be found through 1995. The asterisks next to the t-statistic denote the significance level for a two-tailed test, where * denotes significance at the 10% level, ** denotes significance at the 5% level, and *** denotes significance at the 1% level.

basis. In an attempt to address this issue, Panel B of Table 6 displays results for difference in means and medians tests of the debt ratio of New Zealand firms between the pre- and postimputation periods.8 In all tests, the individual year mean and median and the 1991±1995 average mean and median are compared to the 1985±1988 average mean and median. The debt ratio data for 1991 through 1995 was constructed from responses ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1099

to a survey sent to all firms listed on the New Zealand Stock
Exchange (NZSE) in 1995.9 Two different samples are used in conducting these tests. The first uses all of the firms in a given year for which we have data, which means the same firms are not necessarily studied each year. The second test only uses firms that appeared in our event study sample for which data is available through 1995, which means the same firms are studied each year.
The results shown in Panel B of Table 6 support the event study results displayed in Table 4 and the cross sectional regression results displayed in Panel A of Table 6. When using all available firms, the average mean and median for each year shows a significant decrease at the 5 percent level or better, while the result for the entire 1991±1995 period shows a significant decrease at the 1 percent level. For tests that use only firms that appeared in our event study sample for which we have data through 1995, no significant differences are detected. However, it is important to note that by 1994 and 1995, the debt ratios for our sample firms in the second test (0.37 in both years) have substantially declined below the 1985±1988 debt ratio (0.43). In summary, the results presented in Panel B of Table 6 support the tax shield hypothesis.

5. CONCLUSIONS

This study analyzes shareholder wealth effects associated with five important events concerning the eventual adoption of full dividend imputation in New Zealand by testing the imputation credit and tax shield hypotheses. Using portfolios of firms sorted by five-year average dividend yield and average effective tax rate, event study results were inconclusive regarding the imputation credit hypothesis. However, event study results utilizing portfolios sorted by debt ratio found strong evidence in support of the tax shield hypothesis. These results were confirmed in a crosssectional regression using each firm 's abnormal returns to the five events as the dependent variable. Moreover, difference in means and medians tests that compared the level of debt of New
Zealand firms between the pre- and post-imputation periods reveal that the debt levels of New Zealand firms declined during the period following the adoption of full dividend imputation. ß Blackwell Publishers Ltd 2002

1100

PREVOST, RAO AND WAGSTER

These results support the conclusion that a full dividend imputation system alters the relative attractiveness of debt financing, which suggests that the New Zealand government achieved its goal of reducing the influence of the tax system on the financial policy of firms by adopting full dividend imputation.

APPENDIX

Dates and Announcements
Announcement Date

Announcement

August 20, 1985

The first event was the announcement and release of the `Statement on
Taxation and Benefit Reform. ' This
Statement is the government 's first announcement of its intention to introduce full imputation. The intended reforms to be introduced are detailed along with a variety of other issues, including sales tax, export market development, and development expenditure tax concessions.

April 13, 1987

The second event was a press release by the Minister of Finance. Its intent was to clarify progress on the introduction of the imputation tax system. Finance
Minister Roger Douglas stated that work on the imputation program had been delayed because `other priorities, including work on corporate tax avoidance, were previously hindering work on imputation; however, the pressure of other work has now eased. '

June 18, 1987

The third event was a press announcement by Finance Minister
Douglas that `the Government will ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1101

proceed with the introduction of a full imputation system in 1988/1989 '.
Douglas said that `Treasury is now working on a consultative document on full imputation which I expect to release towards the end of the year. ' He also noted, `We have now had a review of the decision to introduce full imputation. As a result, we have confirmed our decision to proceed [as originally announced]. '
December 17, 1987

The fourth event was the release of the
`Consultative Document on Full
Imputation (December 1987), ' which coincided with a budget press statement by Finance Minister Douglas. He said that
`[t]he consultative committee on full imputation and the international tax regime met for the first time yesterday to plan its agenda. ' The committee was formed to accept feedback from the business community and members of the public regarding The Consultative
Document on Full Imputation, which set out fully the details of the full imputation scheme. Its six chapters total 53 pages and provide an overview of the objectives of the reforms, an analysis of the present tax system, a detailed account of the operation of full imputation, how the scheme will affect shareholders, etc.

April 1, 1988

The fifth event is the date on which imputation formally began. This date is included to detect any reaction associated with the actual changeover to imputation. ß Blackwell Publishers Ltd 2002

1102

PREVOST, RAO AND WAGSTER

NOTES
1

2

3

4

5
6
7

8
9

Other goals were to encourage firms to maintain the corporate form of organization by equalizing taxes on business investment and other forms of income (Statement on Taxation and Benefit Reform, August 20, 1985.), increase capital mobility, disperse ownership of corporate equity, and allow new and rapidly expanding companies to finance investments with new equity rather than having to rely solely on bank debt. The government hoped accomplishing these goals would result in reallocating capital to its most productive use, thus creating higher levels of investment and employment. It is clear, then, that the impact of dividend imputation on firm value is also contingent on the extent to which firms were owned by nonresident shareholders. Fox and Walker (1995) report that the percentage of foreigncontrolled New Zealand companies was only 16.8 percent of all New
Zealand Stock Exchange (NZSE) listed firms in 1985. Since the dates of interest for this study span 1985 to 1988, it is likely that resident New
Zealand taxable shareholders were the primary impetus underlying any associated wealth effects found in this study.
As shown by Howard and Brown (1992), the imputation system is neutral between debt and equity if marginal personal tax rates of taxed domestic investors are less than or equal to the corporate tax rate and the firm is following an optimal dividend policy of paying out the maximum possible imputed dividends. However, the system can actually be biased against debt for personal tax rates greater than the corporate rate.
Because the data for this study comes from two different sources, we are not always able to match historical accounting data for all firms that have returns and dividend data. Specifically, of the 92 firms with dividend and return information that are included in the initial sample, Datex Investor
Services has debt information for 82 firms. Of these 82 firms, 78 have full accounting data.
This definition excludes deferred tax expense, which is consistent with prior research (e.g., Gupta and Newberry, 1997).
From the original 92-firm sample, a complete record of accounting data is available for 78 firms. Thus, all regression models contain 78 observations.
Two other models were tried, and in both cases the debt ratio was negative and significant with no significance being detected for any of the other displayed independent variables. The first alternate model added two interaction variables. The first was to capture effects between low debt and high return on assets firms, and the second was to capture effects between high debt and low return on assets firms. The second alternate model substituted a debt coverage ratio and a dividend coverage ratio for the two interaction variables. The results of these tests are available from the authors upon request.
We wish to thank an anonymous referee for suggesting this test.
For the period 1985 through 1988, responses for 79 out of 234 firms listed on the NZSE were available. For 1991, responses for 74 out of 111 listed firms were available. For 1992, responses for 76 out of 110 listed firms were available. In 1993, responses for 80 out of 125 listed firms were available. In
1994, responses for 86 out of 137 listed firms were available. In 1995, responses for 92 out of 132 listed firms were available.

ß Blackwell Publishers Ltd 2002

DIVIDEND IMPUTATION: THE CASE OF NEW ZEALAND

1103

REFERENCES
Allen, P. and W. Wilhelm (1988), `The Impact of the 1980 Depository
Institutions Deregulation and Monetary Control Act on Market Value and
Risk: Evidence From the Capital Markets ', Journal of Money, Credit and
Banking, Vol. 20, pp. 364±80.
Ashton, D. (1989), `The Cost of Capital and the Imputation Tax System ', Journal of Business Finance & Accounting, Vol. 16, No. 1, pp. 75±88.
Chen, N. and F. Zhang (1998), `Risk and Return of Value Stocks ', Journal of
Business, Vol. 71, pp. 501±35.
Cliffe, C. and A. Marsden (1992), `The Effect of Dividend Imputation on
Company Financing Decisions and Cost of Capital in New Zealand ', Pacific
Accounting Review, Vol. 4, pp. 1±30.
Consultative Document on Full Imputation (December 1987), Wellington, New
Zealand.
Cooper, K., J. Kolari and J. Wagster (1991), `A Note on the Stock Market Effects of the Adoption of Risk-Based Capital Requirements on International
Banks in Different Countries ', Journal of Banking and Finance, Vol. 15, pp.
367±81.
Dann, L. and C. James (1982), `An Analysis of the Impact of Deposit Rate
Ceilings on the Market Value of Thrift Institutions ', Journal of Finance, Vol.
37, pp. 1259±75.
Douglas, Hon. R. O. (1987), Minister of Finance, `Press Release on the
Imputation Tax System ' (April 13).
________ (1987), `Budget Part I: Speech and Annex ' (June 18).
________ (1987), `Press Release on the Tax Consultative Committee '
(December 17).
Eyssell, T. and N. Arshadi (1990), `The Wealth Effect of the Risk-Based Capital
Requirement in Banking: The Evidence from the Capital Market ', Journal of
Banking and Finance, Vol. 14, pp. 179±98.
Fama, E. and K. French (1992), `The Cross-Section of Expected Stock Returns ',
Journal of Finance, Vol. 47, pp. 427±65.
Fox, M. and G. Walker (1995), `Evidence on the Corporate Governance of New
Zealand Listed Companies ', Otago Law Review, Vol. 8, pp. 317±49.
Full Imputation: Report of the Consultative Committee (April 1988), `Consultative
Committee on Full Imputation and International Tax Reform '
(Wellington, New Zealand).
Gupta, S. and K. Newberry (1997), `Determinants of the Variability in Corporate
Effective Tax Rates: Evidence from Longitudinal Data ', Journal of Accounting and Public Policy, Vol. 16, pp. 1±34.
Hamson, D. and P. Ziegler (1990), `The Impact of Dividend Imputation on
Firms ' Financial Decisions ', Accounting and Finance, Vol. 30, pp. 29±53.
Howard, P. and R. Brown (1992), `Dividend Policy and Capital Structure Under the Imputation Tax System: Some Clarifying Comments ', Accounting and
Finance, Vol. 32, pp. 51±61.
International Tax Reform and Full Imputation Part 2, Annex: Draft Legislation. Report of the Consultative Committee (July 1988), Wellington, New Zealand.
James, C. (1983), `An Analysis of Intra-Industry Differences in the Effect of
Regulation ', Journal of Monetary Economics, Vol. 12, pp. 417±32.
Lally, M. (1992), `The CAPM Under Dividend Imputation ', Pacific Accounting
Review, Vol. 4, pp. 31±44.
Monkhouse, P. (1993), `The Cost of Equity Under the Australian Dividend ß Blackwell Publishers Ltd 2002

1104

PREVOST, RAO AND WAGSTER

Imputation System ', Accounting and Finance, Vol. 33, pp. 1±18.
Officer, R. (1994), `The Cost of Capital Under an Imputation Tax System ',
Accounting and Finance, Vol. 34, pp. 1±17.
Prevost, A., R. Rao and M. Hossain (2000), `Board Composition in New
Zealand: An Agency Perspective ', Working Paper (Massey University).
Schwert, W. (1981), `Using Financial Data to Measure Effects of Regulation ',
Journal of Law and Economics, Vol. 24, pp. 121±58.
Statement on Taxation and Benefit Reform (August 20, 1985), Wellington, New
Zealand.
Wagster, J. (1996), `Impact of the 1988 Basle Accord on International Banks ',
Journal of Finance, Vol. 51, pp. 1321±46.
Wood, J. (1997), `A Simple Model for Pricing Imputation Tax Credits Under
Australia 's Dividend Imputation Tax System ', Pacific Basin Finance Journal,
Vol. 5, pp. 465±80.

ß Blackwell Publishers Ltd 2002

References: Allen, P. and W. Wilhelm (1988), `The Impact of the 1980 Depository Institutions Deregulation and Monetary Control Act on Market Value and Ashton, D. (1989), `The Cost of Capital and the Imputation Tax System ', Journal of Business Finance & Accounting, Vol Chen, N. and F. Zhang (1998), `Risk and Return of Value Stocks ', Journal of Business, Vol Cliffe, C. and A. Marsden (1992), `The Effect of Dividend Imputation on Company Financing Decisions and Cost of Capital in New Zealand ', Pacific Consultative Document on Full Imputation (December 1987), Wellington, New Zealand. Cooper, K., J. Kolari and J. Wagster (1991), `A Note on the Stock Market Effects of the Adoption of Risk-Based Capital Requirements on International Dann, L. and C. James (1982), `An Analysis of the Impact of Deposit Rate Ceilings on the Market Value of Thrift Institutions ', Journal of Finance, Vol. 37, pp. 1259±75. Douglas, Hon. R. O. (1987), Minister of Finance, `Press Release on the Imputation Tax System ' (April 13). ________ (1987), `Budget Part I: Speech and Annex ' (June 18). ________ (1987), `Press Release on the Tax Consultative Committee ' (December 17). Eyssell, T. and N. Arshadi (1990), `The Wealth Effect of the Risk-Based Capital Requirement in Banking: The Evidence from the Capital Market ', Journal of Fama, E. and K. French (1992), `The Cross-Section of Expected Stock Returns ', Journal of Finance, Vol Fox, M. and G. Walker (1995), `Evidence on the Corporate Governance of New Zealand Listed Companies ', Otago Law Review, Vol Full Imputation: Report of the Consultative Committee (April 1988), `Consultative Committee on Full Imputation and International Tax Reform ' Gupta, S. and K. Newberry (1997), `Determinants of the Variability in Corporate Effective Tax Rates: Evidence from Longitudinal Data ', Journal of Accounting Hamson, D. and P. Ziegler (1990), `The Impact of Dividend Imputation on Firms ' Financial Decisions ', Accounting and Finance, Vol Howard, P. and R. Brown (1992), `Dividend Policy and Capital Structure Under the Imputation Tax System: Some Clarifying Comments ', Accounting and International Tax Reform and Full Imputation Part 2, Annex: Draft Legislation. Report of the Consultative Committee (July 1988), Wellington, New Zealand. James, C. (1983), `An Analysis of Intra-Industry Differences in the Effect of Regulation ', Journal of Monetary Economics, Vol Lally, M. (1992), `The CAPM Under Dividend Imputation ', Pacific Accounting Review, Vol Monkhouse, P. (1993), `The Cost of Equity Under the Australian Dividend ß Blackwell Publishers Ltd 2002 Officer, R. (1994), `The Cost of Capital Under an Imputation Tax System ', Accounting and Finance, Vol Prevost, A., R. Rao and M. Hossain (2000), `Board Composition in New Zealand: An Agency Perspective ', Working Paper (Massey University). Schwert, W. (1981), `Using Financial Data to Measure Effects of Regulation ', Journal of Law and Economics, Vol Statement on Taxation and Benefit Reform (August 20, 1985), Wellington, New Zealand. Wagster, J. (1996), `Impact of the 1988 Basle Accord on International Banks ', Journal of Finance, Vol Wood, J. (1997), `A Simple Model for Pricing Imputation Tax Credits Under Australia 's Dividend Imputation Tax System ', Pacific Basin Finance Journal,