Harmonization of International Taxation

Managerial Finance
Emerald Article: Harmonization of international taxation: the case of interlisted stocks Steven Graham, Wendy L. Pirie

Article information:
To cite this document: Steven Graham, Wendy L. Pirie, (2003),"Harmonization of international taxation: the case of interlisted stocks", Managerial Finance, Vol. 29 Iss: 1 pp. 33 - 54 Permanent link to this document: http://dx.doi.org/10.1108/03074350310768238 Downloaded on: 24-10-2012 To copy this document: [email protected] This document has been downloaded 409 times since 2005. *

Users who downloaded this Article also downloaded: *
Jon-Arild Johannessen, Bjørn Olsen, (2011),"Aspects of a cybernetic theory of tacit knowledge and innovation", Kybernetes, Vol. 40 Iss: 1 pp. 141 - 165 http://dx.doi.org/10.1108/03684921111117979 James DeLisle, Terry Grissom, (2011),"Valuation procedure and cycles: an emphasis on down markets", Journal of Property Investment & Finance, Vol. 29 Iss: 4 pp. 384 - 427 http://dx.doi.org/10.1108/14635781111150312 John Dorchester Jr, (2011),"Market value, fair value, and duress", Journal of Property Investment & Finance, Vol. 29 Iss: 4 pp. 428 - 447 http://dx.doi.org/10.1108/14635781111150321

Access to this document was granted through an Emerald subscription provided by BIBLIOTECA CENTRALA UNIVERSITARA LUCIAN BL For Authors: If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service. Information about how to choose which publication to write for and submission guidelines are available for all. Please visit www.emeraldinsight.com/authors for more information. About Emerald www.emeraldinsight.com With over forty years' experience, Emerald Group Publishing is a leading independent publisher of global research with impact in business, society, public policy and education. In total, Emerald publishes over 275 journals and more than 130 book series, as well as an extensive range of online products and services. Emerald is both COUNTER 3 and TRANSFER compliant. The organization is a partner of the Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.
*Related content and download information correct at time of download.

Volume 29 Number 1 2003

33

Harmonization of International Taxation: The Case of Interlisted Stocks
by Steven Graham, Department of Economics, College of Arts and Science, Valparaiso University, Valparaiso, Indiana, 46383 and Wendy L. Pirie, College of Business Administration, Valparaiso University, Valparaiso, Indiana, 46383 Abstract The fact that stocks going ex-dividend decline in price by less than the dividend amount is theoretically attributed to the differential taxation of dividend and capital gains or the differential taxation of investor groups. NYSE, Amex and Toronto Stock Exchange listed stocks, and stocks interlisted on these three exchanges, are examined to infer the tax jurisdiction of the marginal investor. The stock price changes relative to the dividends are consistent with a tax clientele effect. Further, the stock price changes are plausible given the tax rates. Ex-dividend day behavior is different for non-interlisted stocks on all three exchanges, suggesting each exchange has a different tax clientele. Canadian firms interlisted on U.S. exchanges exhibit ex-dividend day behavior consistent with the appropriate U.S. exchange’s non-interlisted stocks, suggesting that the marginal investors in these stocks are American. Introduction In this study, the ex-dividend day behavior of interlisted stocks is compared to the exdividend day behavior of stocks that are exclusively listed in the U.S. or in Canada.1 The already high degree of integration of the Canadian and U.S. capital markets, including their stock markets, presents an opportunity to test which country’s tax system influences the price reaction of interlisted stocks on the ex-dividend day. This addresses the question: Does the ex-dividend day behavior of interlisted stocks reflect the stock price behavior of stocks exclusively listed in one country versus the other, or is it an average? Liljeblom, Loflund and Hedvall [2001] find that stocks with high foreign ownership trading on the Helsinki Stock Exchange have ex-dividend day price performance that lies outside the no-arbitrage boundaries for Finnish investors. This is consistent with the results of this paper, that the marginal investor may not be a domestic taxpayer. In making dividend decisions for a corporation, it is necessary that the corporate board be aware of the tax effects on their shareholders. Where a corporation has a significant number of non-resident shareholders, the marginal investor may not be a resident of the country of incorporation. In 1986, significant tax changes occur in both the United States and Canada, increasing the effective tax rate on capital gains income despite a reduction in the tax rate on ordinary income. The relationship between the taxes on dividends and those on capital gains also changes. This paper examines the ex-dividend day behavior of common stocks in the two countries during the period 1982 to 1991, to test the impact of the significant reforms of 1986. McKenzie and Mintz [1992] state, “Although considerable work on the distortions caused by taxes has been undertaken separately in each country, less has been done in the

Managerial Finance

34

way of a systematic comparison of the United States and Canada, especially on a historical basis.”2 Two reasons are given for the importance of this. First, the free-trade agreements potentially impact on capital flows if differences in taxes create differences in costs of production. Second, the tax reforms in both countries have led to claims that the Canadian tax measures are a response to the U.S. measures so that capital flight is prevented. Is this ‘harmonization’ really necessary? If the markets behave as if they are integrated, then perhaps it is not necessary for the government to act to harmonize the tax systems. Increasing integration of global financial markets has resulted in the possibility that stock price behavior is dependent upon institutional details, such as tax regimes, prevalent in a country other than that of incorporation. This is important because it implies that national tax policies may not have the desired incentive effects on publicly traded companies. The marginal investor in a stock may be a non-resident. Bailey [1988], studying Canada’s dual class shares, noted that: “... the characteristics of the marginal investor in the market are not clear, given participation by Canadian, U.S., and other foreign investors with varying tax situations” (p. 1151). Two alternative theories are advanced with respect to ex-dividend day stock price behavior where, commonly, the stock price is observed to decline by less than the amount of the dividend. The change in price, when the stock goes ex-dividend, is asserted to reflect either the marginal tax rate of the marginal investor [Elton and Gruber, 1970] or, if the change in price would be sufficiently different from the amount of the dividend to attract the interest of short-term traders, the transaction costs involved in short-term trading3 [Miller and Scholes, 1982, and Kalay, 1982]. If the empirical research supports either of these two hypotheses, evidence is found that taxes influence behavior in financial markets. Although some researchers have suggested that the ex-dividend day price behavior is not necessarily consistent with a tax clientele effect [Dubofsky, 1992, Skinner, 1993, and Bali and Hite, 1998], Dempsey [2001] concludes that market behavior is consistent with investors capitalizing a firm’s earnings on an after tax basis. Frank and Jagannathan [1998] find that there is an ex-dividend day price anomaly in Hong Kong where there is no taxes on dividends or on capital gains. However, in this paper we find that the marginal investor in a country may not be a taxpayer in that country but may be subject to taxes in another jurisdiction. Koski and Scruggs [1998] find that there is evidence of dividend capture trading but they find little evidence of tax-clientele trading. However dividend capture trading is tax induced trading, where the firm capturing the dividend has different taxation of dividends and capital gains than other investors. In the U.K., Lasfer [1995] finds evidence that taxation does significantly affect ex-dividend day price behavior while Menyah [1993] concludes that the tax-induced dividend clientele effect is not apparent. In New Zealand, Bartholdy and Brown [1999] find a tax clientele effect over the period of their study. New Zealand is particularly suited for studying tax-clientele effects, since corporations may distribute both taxable and non-taxable dividends. In Canada, Athanassakos [1996] finds that the tax clienteles change between corporations and tax-free investors over the period 1970 to 1984 depending on the particular tax regime at the time. Elton and Gruber [1970] find evidence of a dividend clientele: the higher (lower) is the dividend yield, the lower (higher) is the implied tax bracket of the marginal investor. Stockholders in higher tax brackets prefer capital gains to dividends, relative to those in

Volume 29 Number 1 2003

35

lower tax brackets. Kalay [1982, 1984], Lakonishok and Vermaelen [1986], Karpoff and Walkling [1988, 1990] and Stickel [1991] find that short-term trading limits the use of ex-dividend day behavior to infer the existence of dividend clienteles or the marginal tax rates of the marginal investor. Results capture both the impact of the marginal tax rates of the marginal investor and the activity of short-term traders. However, Karpoff and Walkling [1988] find no evidence of short-term trading prior to the introduction of negotiated commissions and support the Elton and Gruber proposition and findings over the period that Elton and Gruber examine. Naranjo, Nimalendran and Ryngaert [2000] confirm that corporate dividend capture programs seem to affect the ex-dividend day prices only after the era of negotiated commissions in 1975 in the U.S. Preceding the work of Elton and Gruber [1970], Campbell and Beranek [1955] and Durand and May [1960] find that, in the U.S., the average stock price decline on the exdividend date is less than the amount of the dividend. They warn that the price drop-off varies widely and that investors cannot necessarily count on an advantage from following policies based on this information, except on average. Heath and Jarrow [1988] further extend the work on risk, stating that the ex-dividend day stock price behavior reflects the equilibrium trading process of investors who require a risk premium. Other empirical studies investigate the effect of a tax regime change on exdividend day behavior [Lakonishok and Vermaelen, 1983, Booth and Johnston, 1984, Barclay, 1987, Grammatikos, 1989, Michaely, 1991, Robin, 1991, Manakyan et al, 1993, and Koski, 1996]. Particularly relevant to this research are the studies on the effect of the Tax Reform Act of 1986 [Dubofsky and Kannan, 1993, Han, 1994, Robin, 1991, and Siddiqi, 1997]. This paper tests the two alternate explanations regarding ex-dividend day price behavior. The dividend distributing common stocks in the sample are divided into groups based on the country of incorporation of the issuing company: Canada or the U.S.; and based on exchange listing: Toronto Stock Exchange (TSE), New York Stock Exchange (NYSE), American Exchange (Amex) or a combination of the TSE and either the NYSE or Amex.4 During the period of the study, 1982 to 1991, significant changes are made in the tax regimes of both countries as well as other institutional changes. The ex-dividend stock price behavior of domestically listed securities (not internationally interlisted) is hypothesized to be different due to substantial differences in each country’s tax regimes. If exclusively Canadian listed, exclusively U.S. listed and interlisted stocks exhibit insignificantly different ex-dividend day behavior, this is evidence of either global integration, that the behavior is expected to be similar or that taxes do not matter. It is possible to distinguish between these explanations on the basis of expected and actual ex-dividend day stock price behavior and by testing for abnormal volume. We find that the ex-dividend day behavior for non-interlisted stocks are different on all three exchanges. This is consistent with each exchange having different marginal investors. However, Canadian firms interlisted on the NYSE or Amex have ex-dividend day behavior consistent with U.S. incorporated NYSE or Amex listed firms, respectively. This suggests that the marginal investor for Canadian firms listed on a U.S. exchange is a U.S. investor. It appears that while domestic tax policy has some impact on ex-dividend day behavior, it is not the only variable with an impact.

Managerial Finance

36

Section II of this paper discusses the institutional changes that occur in the time period under examination, including changes in the tax systems. Section III develops the theory. The model originally developed by Elton and Gruber [1970] is extended and the short-term trading model is considered. Section IV integrates the theory and institutional details to develop specific hypotheses. Section V describes the data while Section VI discusses the methodology. In section VII, the results are presented and the final section summarizes the research and conclusions. II. Institutional Changes II.A. The U.S. During the period 1982 to 1991, two major tax reforms occur, potentially impacting exdividend day behavior. The 1984 Tax Reform Act changes the minimum holding period to be eligible for a dividends received deduction5 from more than 15 days to more than 45 days and disallows certain hedging strategies. These changes are expected to reduce short-term trading as the risk increases with the longer required holding period. The Tax Reform Act (TRA) of 1986 makes sweeping changes. At both the individual and corporate level, the top marginal tax rate on ordinary income is reduced dramatically. Table 1 indicates the effects of the TRA on individual and corporate investors in the top marginal tax brackets. The number of tax brackets for individuals is reduced, from a high in 1984 of 15 to a low in 1991 of 3. As of 1987, all capital gains exclusions are eliminated. The effect is that the rate of tax at the federal level on long-term capital gains increases at the personal and corporate level. The dividends received deduction at the corporate level is also altered such that taxes paid on dividends increase unless the recipient corporation controls at least 80 percent of the dividend-paying subsidiary. As is apparent in Table 3, the 1986 TRA increases the shareholder level taxes on both dividends and capital gains despite reductions in the tax rate on ordinary income. II.B. Canada In Canada, the Federal Budget of 1986 introduces some major tax reforms. Table 2 gives a summary of the tax rates for the years 1982, 1987, 1988 and 1991 for individuals and corporations paying the top marginal tax rates. Tax rates are reduced at both the corporate and the personal level. At the personal level, consistent with tax simplification, the number of tax brackets for personal income tax is reduced from 10 to 3. The capital gains inclusion rate is increased, with a resulting tax increase for capital gains at both the personal and corporate level, despite the tax rate reductions. The dividend tax credit system is modified: the gross-up factor and the dividend tax credit is decreased.6 This increases the effective tax rate on dividends at the personal level for the investor at the top marginal tax rate despite the reduced personal tax rate. The 1986 tax reforms in Canada result in less preferential treatment on income from equity investments, as shown in Table 4. Another change over the period is the institution of negotiable commission rates in Canada effective April 1, 1983. The resulting decline in commission rates makes shortterm trading in Canada more viable.

Volume 29 Number 1 2003
Table 1: Top Marginal Tax Rates in the United States (including surtaxes) Personal - Federal Only (1) 1982 Ordinary income L.T. capital gains
(2)

37

1987 38.5 28.0

1988 28.0 28.0

1991 31.0 28.0

50.0 20.0

Personal- Federal and New York Ordinary income L.T. capital gains
(3),(4)

64.0 25.6

47.3 36.8

36.4 36.4

38.9 35.9

Personal - Federal and California Ordinary income (3) L.T. capital gains 61.0 27.2 47.8 37.3 37.3 37.3 42.0 39.0

Corporate - Federal Only Ordinary income (5),(6) Dividends - domestic L.T. capital gains 46.0 6.9 28.0 40.0 8.0 34.0 34.0 10.2 34.0 34.0 10.2 34.0

Corporate - Federal and New York Ordinary income (3), (5), (7) Dividends - domestic L.T. capital gains 56.0 11.9 38.0 49.0 12.5 43.0 43.0 14.7 43.0 43.0 14.7 43.0

Corporate - Federal and California Ordinary income Dividends - domestic L.T. capital gains Notes:
(1) (3),(5)

55.6 11.9 37.6

49.3 17.3 43.3

43.3 19.5 43.3

43.3 19.5 43.3

The federal rate also represents the combined rate for states with no income tax, such as Texas.

(2)

Ordinary income includes short term capital gains and dividend income from both domestic and foreign corporations. It is assumed that the withholding tax is fully utilized as a foreign tax credit. The state tax is assumed to be calculated on the same taxable income as the federal tax.

(3)

(4)

New York City levies an additional tax at the top rate of 4.3% on individual income. Thus, the top marginal tax rate in New York City on personal income could be as high as 43.2% in 1991. Ordinary income includes short term capital gains.

(5)

(6)

The maximum tax rate on dividends from Canadian issuers equals the tax on ordinary income. Tax relief is provided depending on the level of holdings. It is assumed that at the maximum tax rate, the withholding tax is fully utilized as a foreign tax credit. New York City levies an additional tax at the top rate of 9% on corporate net income. Thus, the top marginal rate in New York City on corporate income could be as high as 52% in 1991.

(7)

Managerial Finance
Table 2: Top Marginal Tax Rates in Canada (including surtaxes) Personal - Federal Only 1982 Ordinary income Dividends-Cdn corp(2) Capital gains
(1)

38

1987 35.0 23.8 17.5 52.0 35.3 26.0 52.1 35.3 26.1 36.6 0.0 18.3 52.1 0.0 26.1 51.6 0.0 25.8

1988 29.9 20.2 19.9 46.1 31.2 30.7 44.6 30.1 29.7 28.8 0.0 19.2 44.3 0.0 29.5 43.8 0.0 29.2

1991 30.5 20.6 22.9 48.0 32.4 36.0 45.2 30.5 33.9 28.8 0.0 21.6 44.3 0.0 33.2 44.2 0.0 33.2

34.0 17.0 17.0 50.3 25.2 25.2 47.1 23.5 23.5 37.8 0.0 18.9 51.8 0.0 25.9 48.8 0.0 24.4

Personal - Federal and Ontario Ordinary income Dividends - Cdn corp Capital gains Ordinary income Dividends - Cdn corp Capital gains Ordinary income Dividends - Cdn corp Capital gain
(3)

Personal - Federal and Alberta

Corporate - Federal Only

Corporate - Federal and Ontario Ordinary income Dividends - Cdn corp Capital gains Ordinary income Dividends - Cdn corp Capital gains Notes:
(1)

Corporate - Federal and Alberta

The ordinary income rate equals the rate on dividends from U.S. corporations. It is assumed that the withholding tax is fully utilized as a foreign tax credit. The Dividends - Cdn corp must be received from a taxable Canadian corporation.

(2)

(3)

Dividends received from a U.S. corporation where holdings are less than 10% are taxed as ordinary income with a foreign tax credit available for withholding taxes. Where holdings are at least 10%, there are no Canadian income taxes imposed but the foreign tax credit for withholding taxes cannot be applied. In these cases, the effective tax rate equals the withholding tax rate.

III. Theory The tax clientele model assumes that investors with different tax circumstances invest in securities with different characteristics. Given this model, the ex-dividend day behavior of stock prices can be used to find the implied relationship between the tax on dividends and the tax on capital gains for the marginal investor in the stock [Elton and Gruber, 1970]. Elton and Gruber consider an owner who has decided to sell the stock. Before or on the final cum-dividend date, i.e., the trading day immediately preceding the day the stock

Volume 29 Number 1 2003
æ 1- td ö ÷ Table 3: Indifference Tax Factorsç ç ÷for the U.S. è 1- tc ø For the investor in the top marginal tax bracket. Individual Federal Combined- New York Combined- California Corporate Federal Combined- New York Combined- California 1.293 1.421 1.338 1.293 1.421 1.338 1.293 1.421 1.338 1.293 1.421 1.338 1.293 1.421 1.338 1.394 1.535 1.459 1.361 1.496 1.420 1.361 1.496 1.420 1.361 1.496 1.420 1982 .625 .484 .536 1983 .625 .484 .536 1984 .625 .484 .536 1985 .625 .486 .536 1986 .625 .489 .536 1987 .854 .834 .833 1988 1.0 1.0 1.0 1989 1.0 1.0 1.0 1990 1.0 1.0 1.0

39

1991 .958 .953 .951 1.361 1.496 1.420

æ 1- td ö ÷ Table 4: Indifference Tax Factors ç ç ÷ for Canada è 1- tc ø For an investor in the top marginal tax bracket. Individual Federal Combined- Ontario Combined- Alberta Corporate Federal Combined- Ontario Combined- Alberta 1.233 1.350 1.323 1.227 1.351 1.316 1.233 1.359 1.323 1.220 1.342 1.307 1.233 1.364 1.323 1.224 1.353 1.348 1.238 1.418 1.412 1.238 1.418 1.412 1.276 1.497 1.488 1.276 1.497 1.497 1982 1.0 1.0 1.0 1983 1.0 1.0 1.0 1984 1.0 1.0 1.0 1985 1.0 1.0 1.0 1986 1.0 1.0 1.0 1987 .924 .874 .876 1988 .996 .993 .994 1989 .996 .993 .994 1990 1.028 1.054 1.051 1991 1.030 1.056 1.051

goes ex-dividend (hereafter, the cum-date), the decision faced is to sell on the cum-date and not receive the dividend or to sell on the ex-dividend day (hereafter, the ex-date) and receive the dividend. If the stock is sold cum-dividend, the after-tax value received is PX -1 - t c ( PX -1 - PO ) Where PX-1 is the selling price on the cum-date, PO is the adjusted cost base in the stock for capital gains purposes, and tc is the marginal tax rate on capital gains. is If the owner waits to sell on the ex-date, the after-tax value expected to be received D (1- t d ) + E[PX ]- t c ( E[PX ]- PO ) where D is the amount of the cash dividend, td is the marginal tax rate on dividends, and E[PX] is the expected price on the ex-date. (2) (1)

Managerial Finance
when

40

A risk-neutral owner is indifferent between selling on the cum-date and the ex-date (3)

PX -1 - t c ( PX -1 - PO ) = D (1- t d ) + E[PX ]- t c ( E[PX ]- PO ) This can be rearranged to obtain PX -1 - E[PX ] 1- t d = D 1- t c

(4)

Transaction costs are not included since they are assumed to be equal in both cases. This model also assumes that all cash flows occur at the same time, that is, there is no discounting. Likewise, considering a buyer instead of a seller and assuming no discounting, we also get Equation (4). 1- t d is called the indifference tax factor (ITF) in this paper. This ITF 1- t c varies by investor group and with changes in the tax regime over time. The ITF reflects the amount a capital gain must be for an investor to be indifferent between it and a onedollar dividend. If the market is in equilibrium, the marginal investor is indifferent between selling on the cum-date or the ex-date. The term The empirical testing of the model is restricted to the left-hand side of Equation (4) that is the same across all models. In interpreting the empirically estimated indifference tax factors, the right hand sides of the various models are considered which represent the theoretically estimated indifference tax factors. These estimates do not significantly differ between the models for a seller with and without discounting. The model for a longterm buyer with discounting lowers the theoretically estimated indifference tax factors. III.B. Short-term trading Short-term traders transact in the stocks because of the dividends, as opposed to other investors who are interested in the stock as part of a longer-term strategy. Short-term traders plan on holding the stock only long enough to benefit from the transaction as a result of their particular tax situation. This holding period varies across investor groups and between countries. If the decline in stock price varies from the dividend amount, short-term traders with an indifference tax factor of 1.000 will enter the market to obtain arbitrage profits and will drive the implied ITF towards 1.000. Further, short-term trading by corporate shareholders that prefer dividends to capital gains impacts on the ITF. Due to transactions costs, this process is somewhat hampered and the observed ITF can vary from 1.000 even given the activity of short-term traders. To distinguish between ITF’s driven by short-term trading versus those resulting from the activities of long-term traders, consistent with the tax-clientele model, volume around ex-dividend days is investigated [Lakonishok and Vermaelen, 1986, Stickel, 1991, Shaw, 1991, and Manakyan et al. 1993]. Evidence of no abnormal volume is consistent with the long term trading hypothesis. Evidence of abnormal volume cannot be so easily interpreted. Campbell and Beranek [1955] suggest strategies based on tax circumstances if a transaction is already being contemplated. A strict application of such strategies could lead to clustering around the

Volume 29 Number 1 2003

41

ex-date by long-term investors and abnormal volume.7 However, this clustering would be the result of postponement of activity until the ex-date and acceleration of activity to the cum-date. Therefore, unlike the case where there are short-term traders, the cumulative volumes for a period around the ex-date should not be abnormal, as other days are expected to exhibit lower than normal volumes in order for the activity to cluster around the ex-date. Short-term traders transact in stocks where the expected profits are greatest. Shortterm traders are expected to be more active in stocks with a low relative bid-ask spread, low transaction costs and greater liquidity. These factors are proxied by the logarithm of price [Demsetz, 1968, Tinic and West, 1972, Benston and Hagerman, 1974, Branch and Freed, 1977, Edmister, 1978, Walkling and Edmister, 1983, and Stoll and Whaley, 1983], the standard deviation of price [Walkling and Edmister, 1983], and the daily trading value [Demsetz, 1968, Tinic and West, 1972, Benston and Hagerman, 1974, Branch and Freed, 1977, Walkling and Edmister, 1983, and Stoll, 1989]. The hypothesis is that, as liquidity increases, as the price variability decreases, and/or as the relative bid-ask spread decreases, abnormal volume will increase, reflecting the increased participation of short-term traders. Further, the indifference tax factor is expected to be closer to one as the abnormal volume increases, reflecting short-term trading activity. IV. Hypotheses This section begins with a discussion of the indifference tax factors (ITFs) for various investor groups calculated from their tax rates. Arbitrage opportunities result from these differential ITFs.8 Then, specific hypotheses are presented which integrate the institutional detail and the theory. IV.A. Indifference Tax Factors The ITFs reflect how large a capital gain must be for an investor to be indifferent between it and a dividend of one dollar. An ITF less (greater) than 1.0 indicates a preference for capital gains (dividends). Tables 3 and 4 show the ITFs for individual and corporate investors in the top marginal tax brackets in the United States and Canada respectively for 1982 to 1991. In the remainder of this section, unless otherwise specified, individual investors and corporate investors refer to those investors in the top marginal tax brackets. In the U.S., the lower bounds on ITFs are those of individual investors in the top marginal tax bracket and the upper bounds are the ITFs of corporations with holdings of at least 80% in the dividend-paying corporation. However, when ownership levels reach 80%, policies are set by the owning corporation and the stock is often thinly traded. The owner corporation, generally, is not active in the market with respect to this stock and is not the marginal investor. It is reasonable to assume the upper bounds are the ITFs of corporate investors in the top marginal tax bracket. Individual investors’ ITFs before 1987 indicate a strong preference for capital gains. In 1987, the preference is reduced. In 1988 to 1990, the ITFs show indifference between capital gains and dividends. In 1991, as the result of capping the tax on capital gains, a slight preference for capital gains is restored.

Managerial Finance

42

Corporate investors’ ITFs indicate a strong preference for dividends over the entire period. This level of preference increases as the level of ownership goes to at least 20% for years subsequent to 1987 and increases even further as ownership reaches at least 80%. The gap between the ITFs of the individual and corporate investors is approximately 0.67 for 1982 to 1986, 0.54 for 1987 and 0.36 for 1988 to 1991, at the federal level. Introducing state taxes results in a widening of the gap. With combined federal and New York taxes, the gap is as wide as approximately 0.94 for 1982 to 1986. Over the same period, considering California taxes, the spread is approximately 0.80. In Canada, the ITFs for the individual and corporate investors represent the extreme ranges of ITFs for 1982 to 1989. For 1990 and 1991, the lower boundaries are 1.00, the ITFs of investors who are either tax-exempt or pay taxes on dividends and capital gains at the same tax rate. Individual investors in the top marginal tax bracket are indifferent between the receipt of a dividend of one dollar or a capital gain of one dollar during the period 1982 to 1986 inclusive. In the years 1987 to 1989, a slight preference is indicated for capital gains by these investors and in the final years of study, 1990 to 1991, a slight preference is indicated for dividends. Individual investors in lower marginal tax brackets prefer dividends in all years. The disparity between the ITFs of individual investors in the top marginal tax bracket and those in lower marginal tax brackets decreases over the years. Corporate investors prefer dividends from a Canadian corporation to capital gains for the entire period. The gap between the ITFs of individual and corporate investors is approximately 0.23 for 1982 to 1986, 0.30 for 1987 and 0.25 for 1988 to 1991 at the federal level. Using a combined federal and Ontario tax rate, the gap is approximately 0.35 for 1982 to 1986, 0.48 for 1987 and 0.43 for 1988 to 1991. A combined federal and Alberta rate leads to similar results. The differences between the ITFs of the different investor groups create the possibility of arbitrage opportunities, commonly referred to as dividend capture programs. Investors will engage in arbitrage opportunities to the point where the costs equal the profits. The introduction of transactions costs raises the price decline required by shortsellers and reduces the price decline required by purchasers. If short-term trading does occur, it will be in stocks with low relative bid-ask spreads, low transaction costs and high liquidity. IV.B. Specific Hypotheses Hypothesis 1: The indifference tax factor is positively related to the dividend yield. As the dividend yield increases, the indifference tax factor increases. Higher-yielding stocks are held by investors with a preference for dividends as a result of preferential tax treatment on those. Investors ignore dividend yield and select portfolios that are efficient on a before-tax basis. Risk cannot be held constant in the process of altering portfolios for higher after-tax returns.

The null hypothesis:

Volume 29 Number 1 2003

43

No difference exists in the indifference tax factor for different dividend yields. Hypothesis 2: The ITFs comparing Canada and the U.S., where there is no evidence of short-term trading, are different for each of the years 1982 to 1986 inclusive, and not different for the years 1987 to 1991 inclusive. The ITFs on interlisted stocks are not different from those in the U.S. regardless of the country of incorporation. This would be consistent with the marginal investor for Canadian stocks interlisted on a U.S. exchange being a U.S. resident taxpayer.

Hypothesis 3:

V. Data Common stocks that pay cash dividends, ordinary and/or extraordinary, with ex-dividend dates after January 2, 1982 and before January 1, 1991 and are listed on the NYSE, the Amex or the TSE represent the entire potential sample. If a stock is listed on the TSE and is interlisted on the NASDAQ or any U.S. exchange but not the NYSE or Amex for some period, it is excluded from the sample over that period. Any common stocks where the issuer is not incorporated in Canada or in the United States are separately identified. The data files from the Center for Research in Security Prices (CRSP) are the primary source of data for the NYSE, Amex and all interlisted stock listings. The data files from the Canadian Financial Markets Research Center (CFMRC) are used for the TSE listings. If a firm has an ordinary and an extraordinary cash dividend paid on the same exdate, the two dividends are added together and treated as one dividend distribution. Dividends reported in foreign currencies are converted to the currency of the market using the exchange rate for large transactions as reported in the Federal Reserve bulletin. If there is another distribution event, such as a stock dividend, stock split, or spinoff, on the same date or within twenty trading days of the cash dividend ex-date, that exdate is excluded from the sample for that firm, to remove coincident events which may affect the valuation of the firm’s stock or its trading volume. If there is no valid transaction price on the ex-date or on the cum-date, that observation is dropped from the sample. Although CRSP reports a return in such cases, the return is based on the average of the bid and the ask price, which may not be representative of a transaction price. If there is insufficient data for estimating the market model parameters, the observation is excluded from the sample. All of the cash dividend distributions are identified for the three exchanges. The NYSE, Amex and TSE firms identified as distributing cash dividends are coded to identify their status based on country of incorporation and exchange listing. There are nine possible circumstances based on primary listing location (which coincides with country of incorporation)/secondary listing location: exclusively NYSE (36,058 observations), NYSE/TSE (1,171), Other Foreign/NYSE (241), exclusively Amex (7,881), Amex/TSE (7), Other Foreign/Amex (101), exclusively TSE (5,413), TSE/NYSE (550), and

Managerial Finance

44

TSE/Amex (441). The sample of NYSE and Amex listed stocks, including interlisted stocks, is 46,450 observations and the TSE sample is 5,413 observations. VI. Methodology The indifference tax factors computed on ex-dividend dates are hypothesized to be due to the different marginal tax rates on dividend income and capital gains (losses). In general, the ITF is computed as $ PX -1 - P 1- t d X = D 1- t c where td is the marginal tax rate on dividends, tc is the marginal tax rate on capital gains, PX-1is the closing price of the stock on the cum-day, the day prior to going ex-dividend, $ is the closing price on the ex-dividend date,9 adjusted for the market reP X turns as discussed below, and D is the value of the cash dividend distributed on the dividend pay date. This study adjusts the closing prices using a risk adjusted market model: $ = P X where PX 1+ {R f + b i ( R M - R f )} PX is the closing price on the ex-date, RM is the return to the market on the ex-date, bi is the systematic risk on stock i estimated over the estimation period, and Rf is the daily return on 90 day Treasury bills for the U.S. or Canada, depending on the exchange, adjusted from a discount basis. This adjustment results in a negative return in some cases where the market drops overnight but begins with the assumption of positive expected returns. The return to the equally weighted CRSP index, including dividends, is used for stocks trading on the NYSE or Amex for RM. The return to the TSE 300 index, including dividends, as computed by the CFMRC, is used for for non-interlisted stocks on the TSE. The market returns are with dividends, since the stock of interest is accruing the dividend between ex-dividend dates. If market returns without dividends are used, the market returns would be understated, since the value of stocks drops on the ex-dividend day, biasing the market returns down. (6) (5)

Volume 29 Number 1 2003

45

For observations from the CRSP files, bi is computed using the methods developed in Scholes and Williams [977) to adjust for nonsynchronous trading. For observations from the CFMRC files, bi is that calculated by CFMRC. The estimation period is the interval beginning (ending) 60 trading days before (after) the ex-dividend date. If there is a cash dividend distributed within this period, then the beginning (ending) date is two trading days after (before) this prior (subsequent) distribution.10 The ex-dividend date and the two prior and the five subsequent trading days are excluded in all cases. The maximum estimation period for a stock is 113 days. The five subsequent trading days are excluded in order that the days up to and including the record date are excluded. The possibility of special settlements creates the necessity for this. If there is no valid closing price available for a day, that day and the subsequent day are also excluded. The returns for both these days are computed based on the average of the closing bid and ask prices, not on closing transaction prices and as such may be inaccurate. Eliminating such data ensures that the actual returns data used are based on trade data only. If less than 40 valid returns are available during the estimation period, that observation is excluded. If more than 5 of the 11 trading days centered on the ex-date lack valid prices, that observation is excluded. VI.A. Adjustment for Heteroscedasticity The ITFs, as computed, are unbiased and normally distributed with variance conditional on the variance of the prices for each firm over the estimation period, the amount of the dividend and the expected return to the stock. As developed in Michaely [1991], the variance of the ITFs is conditional on the dividend yield (see Michaely, Appendix A, page 857) and the variance of the returns over the estimation period. The slight difference in the two approaches arises from the assumption as to what the stochastic variable is. Michaely assumes the returns are uncertain, while this paper uses the ex-date price as the uncertain variable, from which the uncertain return is derived. Also, this paper adjusts the ex-date price for the expected overnight return on the stock. To control for the conditional heteroscedasticity and to obtain efficient estimators of the ITFs, generalized least squares is used where the mean ITFs are weighted by the variance of the prices and inversely weighted by the dividend and one plus the expected return. The mean ITF for a portfolio of firms and ex-dividend dates is estimated by ITF = å
i=1 N

ITFi s2 Pi {D * (1+ E[R i ])}2 i

å {D
j =1

N

s2 Pj
* j

(1+ E[R j ])}2

(7)

where

is the variance of the price of stock i over the estimation period, s2 Pi is the discounted value of the dividend on stock i from the pay-date to the D* i cum-date using the daily return on 90 day T-bills for the U.S. or Canada, and E[Rj] is the expected return on stock i.11

Managerial Finance

46

The result is GLS estimates of the ITFs for the portfolios of firms and dates. Gagnon and Suret [1991] find that the power of the models used in prior research is inadequate to find significant results, since the samples are often small. By using GLS, the power of the tests is improved and their concern is addressed. The entire sample is divided into quintiles by dividend yields. This results in groups of less than equal size by year or exchange but allows for comparability across exchanges and years.12 Many studies have reported their findings by deciles but the use of quintiles increases cell sizes and still allows for comparison of results to those of prior research. The GLS estimates of ITFs are compared by quintiles. This approach is similar to that of previous research. VII. Results The conclusions are structured to follow the hypotheses developed in Section IV. Also, when the results differ based on the country of incorporation or exchange listing, this is noted. Unless otherwise specified, the two methods have similar results and the results are similar across listing types and countries of incorporation. Hypothesis 1: The indifference tax factor is positively related to the dividend yield. Due to the clientele effect, as the dividend yields increase the indifference tax factors increase. Using GLS, Hypothesis 1 is supported for U.S. non-interlisted stocks (Table 5, rows NYSE/NYSE and Amex/Amex). Hypothesis 1 is not supported for TSE noninterlisted stocks (Table 5, row TSE/TSE). In addition, using GLS, the null hypothesis that the mean ITFs equal 1.0, is rejected for U.S. non-interlisted stocks but cannot be rejected for TSE non-interlisted stocks. The ITFs are consistent with the ITFs calculated in Tables 3 and 4. Using regression, Hypothesis 1 is supported (Table 6). The natural logarithm of dividend yield is significantly positively related to the indifference tax factor over all observations.13 However, it is found upon further examination of the observations from the CRSP files that the quintile with the lowest yield generates a significantly different model than the top four yield quintiles. In fact, in this quintile, yield is negatively related to the indifference tax factor. Low dividend stocks as opposed to low yield stocks are excluded in some prior research. Hypothesis 2, that indifference tax factors, where there is no evidence of short-term trading, are different for U.S. and Canadian incorporated non-interlisted firms, is supported. The null hypothesis, that they are equal, is rejected at the 5 percent level of confidence. See Table 5, rows T/T = N/N and T/T = A/A which compares the ITFs for stocks which are only listed on the TSE with stocks that are only listed on the NYSE or the Amex respectively. Hypothesis 3, that the indifference tax factors for interlisted stocks are equal to those in the U.S., cannot be rejected at the 5 percent level of confidence using GLS. Using regressions, we conclude that the country of incorporation significantly affects the observed indifference tax factors for stocks listed on the NYSE, while firms traded on the Amex demonstrate no significant difference, despite different countries of incorporation (Table 6, row TSE/NYSE and row TSE/Amex).

Volume 29 Number 1 2003

47

Table 5: Mean ITFs by Primary Listing/Secondary Listing and Probabilities the Mean ITFs are Equal Primary/ Secondary Listing Yieldq 1 Yieldq 2 Yieldq 3 Yieldq 4 Yieldq 5

Mean ITFs TSE/NYSE NYSE/ NYSE NYSE/TSE TSE/TSE TSE/Amex Amex/ Amex 1.183 1.023 .996 .076 .364 .890 .467 .749 .518 .588 1.132 .830 Probabilities N/T=N/N T/N=N/N T/A=A/A N/T=T/N T/N=T/A T/T=N/N T/T=A/A N/T=T/T T/N=T/T T/A=T/T .9210 .5741 .1341 .6319 .0659 .0001
*

.840 .806 .836 .483 .603 .861

.932 .863 .921 .443 .797 .880

.908 1.108 1.031 .668 1.195 .993

.4266 .3646 .3375 .9038 .1263 .0991 .0351* .8170 .7071 .0883

.8743 .9061 .3512 .9924 .5404 .0015
*

.7419 .8082 .8098 .9750 .7573 .0001
*

.6636 .5626 .7110 .7480 .6509 .0001* .0236* .0629 .4993 .3299

.0001* .0015* .0002* .4248

.0024* .0928 .2294 .6703

.0006* .0142* .0968 .3081

Since there are only 7 observations available for Amex/TSE, it is not presented. For example: T/N=N/N shows the probability that the ITF for the TSE/NYSE equals the ITF for the NYSE/NYSE. * Significant at a 5 percent level of confidence.

In conclusion, the indifference tax factors, calculated based on ex-dividend day behavior of common stocks, demonstrate some behavior consistent with a tax effect and are reasonable compared to the theoretical indifference tax factors developed in Section IV. However, some of the differences in indifference tax factors seem to be attributable to causes other than taxes. Canadian firms interlisted on the NYSE or Amex have exdividend day behavior consistent with U.S. incorporated NYSE or Amex firms, respectively. This suggests that the marginal investor for Canadian firms is a U.S. investor if the stocks are listed on a U.S. exchange.

Managerial Finance
Table 6: Regression Results: Incorporation and Exchange Listing Intercept Intercept ln (Yield) Abnormal Volume TSE/NYSE TSE/Amex NYSE/TSE NYSE/NYSE Amex/Amex Foreign/NYSE Foreign/Amex Adjusted R2 N SSE TSS F-Statistic 0.05691 (2.108) -0.35042 (-3.415) 0.08380 (1.494) 0.0094 37154 42841.98663 43257.30348 45.011 1.0897 (31.389) 0.05701 (8.219) 0.00333 (13.926) -0.09372 (-2.148) 0.07883 (1.152) 0.07904 (2.096) 0.05701 (8.219) 0.00333 (13.926) 0.99596 (20.750) 1.16851 (14.518) 1.16872 (25.664) 1.23524 (44.277) 1.14659 (39.432) 0.73926 (6.077) 1.17348 (18.236) 0.5400 37154 42841.98663 93153.89181 4846.844 No Intercept

48

The dependent variable is the estimated ITF for each observation. Each cell contains the estimated coefficient and the t-statistic that the coefficient could be zero. For the model with an intercept, the stocks with their home market as NYSE and trading on NYSE are used as the No Dummy case. The coefficients show the difference from the No Dummy observations. The R2, F-statistic and TSS for the two models are not comparable since the No Intercept model is uncentered.

VIII. Summary and Conclusions Evidence is found of a tax clientele effect in that, as the dividend yields increase, the indifference tax factors increase. This result is consistent with the Elton and Gruber proposition that the marginal investor can be inferred by examining ex-dividend day behavior. However, the tax changes of 1986 do not lead to the hypothesized change in the indifference tax factors in the U.S. The indifference tax factors for non-interlisted stocks are different on all three exchanges. This is consistent with each exchange having different marginal investors. However, Canadian firms interlisted on the NYSE or Amex have ex-dividend day behavior consistent with U.S. incorporated NYSE or Amex listed firms, respectively. This suggests that the marginal investor for Canadian firms listed on a U.S. exchange is a U.S. in-

Volume 29 Number 1 2003

49

vestor. It appears that while domestic tax policy has some impact on ex-dividend day behavior, it is not the only variable with an impact. For a corporation making dividend policy decisions, it is necessary to consider the tax position of the shareholders. A firm that makes a dividend decision inconsistent with its shareholder clientele’s tax position will cause considerable turnover in its shareholders. To complicate matters, a corporation cannot assume that its shareholder clientele are residents of the country of incorporation. Where a corporation has significant nonresident shareholders, their tax position must also be considered. Future research might investigate the reason for the significant difference between the ex-dividend day behavior of NYSE and Amex listed stocks. Acknowledgements The authors gratefully acknowledge the helpful suggestions received when an earlier version of this paper was presented at the 2nd Annual Conference on Multinational Financial Issues. We gratefully acknowledge the suggestions and comments of participants at workshops at Lehigh University and at the University of Mississippi.

Managerial Finance
Endnotes

50

1. In this paper, interlisted refers to stocks that are incorporated in either Canada or the U.S., listed on an exchange in the country of incorporation and also listed on an exchange in the other country. Interlisted stocks account for 28.2%, 13.3% and 4.6% of the trading volume per year respectively over the period, 1982 to 1991, on the TSE, Amex and NYSE respectively. 2. McKenzie and Mintz [1992] Tax Effects on the Cost of Capital in Canada-U.S. Tax Comparisons, Chicago: University of Chicago Press, p.189. 3. The term short-term trading refers to trading for dividend capture and is unrelated to the tax code that defines short-term versus long-term trading for tax purposes. 4. No stocks are interlisted between the NYSE and Amex. 5. The dividends received deduction refers to the portion of dividends received by a U.S. corporation that is excluded from taxation. 6. Canada uses a modified imputation tax system to reduce the impact of double taxation on dividend distributions to individuals. Dividends are paid from corporate income on which taxes have already been levied but individuals are allowed a tax credit on dividend income to offset the corporate taxation on the dividends paid. 7. Stickel [1991] p.48 states “Incentives for long-term traders to time their trades and for short-term traders to capture dividends both increase with dividend yield. Thus, a positive relation between ex-day abnormal volume and dividend yield cannot establish the existence of dividend capture.” 8. Heath and Jarrow [1988] correctly point out that the term arbitrage opportunities is applied loosely in a large amount of the literature on ex-dividend day behavior. A true arbitrage opportunity involves no risk, which is not true in these cases. 9. Closing prices are used instead of opening prices for consistency and for a valid comparison to the market returns. See Stickel [1991] for a discussion of the difficulties in using opening prices. Given the exchange rules regarding the adjustment of limit orders to buy and stop orders to sell, the use of opening prices may potentially bias the results. 10. Stickel [1991] uses an estimation period 75 days on each side of the ex-dividend date, which will include the adjoining dividend dates, since quarterly dividends are distributed every 63 trading days, on average. Michaely [1991] uses an estimation period 25 days on each side of the ex-dividend date, which is smaller than used here. 11. E [Ri] is computed as Rt +bi (E [RM]-Rf ) where the E [RM] is estimated to be 0.00048 per calendar day.

Volume 29 Number 1 2003

51

12. Shifts occur in these averages over time as a result of inflation, growth and institutional changes. Papaiannou and Savarese [1994] document changes in corporate dividends as a result of the 1986 TRA. These changes result in a shift in average dividend yields based on quintiles: the average dividend yield in the top quintile is not necessarily the same in 1988 as in 1986. The use of quasi-quintiles developed on the basis of the actual quintiles for the entire period 1982 to 1991, results in groups of different size for different years but with similar average dividend yields or other variable of interest across years. 13. Testing confirms that the natural logarithm of the dividend yield is a better variable than using the natural logarithms of price and dividend separately.

Managerial Finance
References

52

Athanassakos, George. Tax induced trading volume around ex-dividend days under different tax regimes: The Canadian Experience 1970-1984. Journal of Business Finance and Accounting 23(4), 1996, 557-584. Bailey, Warren. Canada’s Dual Class Shares: Further Evidence on the Market Valuation of Cash Dividends. Journal of Finance 43(5), 1988, 1143-1160. Bali, Rakesh, and Hite, Gailen L. Ex dividend day stock price behavior: discreteness or tax-induced clienteles? Journal of Financial Economics 47(2), 1998, 127-159. Barclay, Michael J. Dividends, Taxes, and Common Stock Prices: The Ex-Dividend Day Behavior of Common Stock Prices before the Income Tax. Journal of Financial Economics 19(1), 1987, 31-44. Bartholdy, Jan, and Brown, Kate. Ex-dividend day pricing in New Zealand. Accounting and Finance 39(2), 1999, 111-112. Benston, G.J. and Hagerman, E.L. Determinants of Bid-Ask Spreads in the Over-theCounter Market. Journal of Financial Economics 1, 1974, 353-364. Booth, L. D. and Johnston, D. J. The Ex-Dividend Day Behavior of Canadian Stock Prices: Tax Changes and Clientele Effects. Journal of Finance 39(2), 1984, 457-476. Branch, B. and Freed, W. The Impact of Option Listing on the Price and Volume of the Underlying Stock. Financial Review 16, 1977, 1-15. Campbell, James A. and Beranek, William. Stock Price Behavior on Ex-Dividend Dates. Journal of Finance 10(4), 1955, 425-429. Dempsey, Mike. Investor tax rationality and the relationship between dividend yields and equity returns: An explanatory note. Journal of Banking and Finance 25(9), 2001, 1681-1686. Demsetz, H. The Cost of Transacting. Quarterly Journal of Economics 82, 1968, 33-53. Dubofsky, David A market microstructure explanation of ex-day abnormal returns. Financial Management 21, 1992, 32-43. Dubofsky, David, and Kannan, Srinivasan. Hedged dividend capture and ex-day returns: an empirical update. Journal of Business Finance and Accounting 20(5), 1993, 725-735. Durand, David. and May, Alan M. The Ex-Dividend Behavior Of American Telephone And Telegraph Stock. Journal of Finance 15(1), 1960, 19-31. Edmister, Robert O. Commission Cost Structures: Shifts and Scale Economies. Journal of Finance 33, 1978, 477-486. Elton, Edwin J. and Gruber, Martin J. Marginal Stockholder Tax Rates and The Clientele Effect. Review of Economics and Statistics 52(1), 1970, 68-74.

Volume 29 Number 1 2003

53

Frank, Murray, and Jagannathan, Ravi. Why do stock prices drop by less than the value of the dividend? Evidence form a country without taxes. Journal of Financial Economics 47(2), 1998, 161-188. Gagnon, Jean-Marie and Suret, Jean-Marc. Ex-Dividend Day Price Changes and Implied Tax Rates: An Evaluation. Journal of Financial Research 14(3), 1991, 255-262. Grammatikos, Theoharry. Dividend Stripping, Risk Exposure, and The Effect of the 1984 Tax Reform Act on the Ex-Dividend Day Behavior. Journal of Business 62(2), 1989, 157-174. Han, Ki C. The effect of the 1986 Tax Reform Act on ex-dividend day return behavior. Journal of Financial Research 17(2), 1994, 175-185. Heath, David C. and Jarrow, Robert A. Ex-Dividend Stock Price Behavior and Arbitrage Opportunities. Journal of Business 61(1), 1988, 95-108. Kalay, Avner. The Ex-Dividend Day Behavior of Stock Prices: A Re-Examination of the Clientele Effect. Journal of Finance 37(4), 1982, 1059-1070. Kalay, Avner. The Ex-Dividend Day Behavior of Stock Prices: A Re-Examination of the Clientele Effect: A Reply. Journal of Finance 39(2), 1984, 557-561. Karpoff, Jonathan M. and Walkling, Ralph A. Short-Term Trading around Ex-Dividend Days: Additional Evidence. Journal of Financial Economics 21(2), 1988, 291-298. Karpoff, Jonathan M. and Walkling, Ralph A. Dividend Capture in NASDAQ Stocks. Journal of Financial Economics 28(1/2), 1990, 39-66. Koski, Jennifer. A microstructure analysis of ex-dividend stock price behavior before and after the 1984 and 1986 tax reform acts. Journal of Business 69, 1996, 313-338. Koski, Jennifer, and Scruggs, John T. Who trades around the ex-dividend day? Evidence from NYSE audit file data. Financial Management 27(3), 1998, 58. Lakonishok, Josef and Vermaelen, Theo. Tax Reform and Ex-Dividend Day Behavior. Journal of Finance 38(4), 1983, 1157-1179. Lakonishok, Josef and Vermaelen, Theo. Tax-Induced Trading around Ex-Dividend Days. Journal of Financial Economics 16(3), 1986, 287-319. Lasfer, M. Ameziane Ex-Day Behavior: Tax or Short-Term Trading Effects. Journal of Finance 50(3), 1995, 875-897. Liljeblom, Eva, Loflund, Anders and Hedvall, Kaj. Foreign and domestic investors and tax induced ex-dividend day trading. Journal of Banking and Finance 25(9), 2001, 1687-1716. Manakyan, Herman, Liano, Kartono and Huang, Gow Cheng. The 1986 Tax Reform Act and Ex-Dividend Day Price and Volume Patterns: Evidence from Organized Exchanges and The OTC Market. Journal of Economics and Finance 17(1), 1993, 135-148.

Managerial Finance

54

McKenzie, Kenneth J. and Mintz, Jack M. Tax Effects on the Cost of Capital. In Canada-U.S. Tax Comparisons, ed. Shoven, John B., and Whalley, John. Chicago: University of Chicago Press, 1992, 189-216. Menyah, Kojo. Ex-dividend equity pricing under UK tax regimes. Journal of Business Finance and Accounting 20(1-2), 1993, 61-81. Michaely, Roni. Ex-Dividend Day Stock Price Behavior: The Case of the 1986 Tax Reform Act. Journal of Finance 46(3), 1991, 845-860. Miller, Merton H. and Scholes, Myron S. Dividends and Taxes: Some Empirical Evidence. Journal of Political Economy 90(6), 1982, 1118-1141. Naranjo, Andy, Nimalendran, M., and Ryngaert, Mike. Time Variation of Ex-Dividend Day Stock Returns and Corporate Dividend Capture: A Reexamination. Journal of Finance 55(5), 2000, 2357-2372. Papaioannou, George J. and Savarese, Craig M. Corporate Dividend Policy Response to the Tax Reform Act of 1986. Financial Management 23(1), 1994, 56-63. Robin, Ashok J. The Impact of the 1986 Tax Reform Act on Ex-Dividend Day Returns. Financial Management 20(1), 1991, 60-70. Scholes, Myron and Williams, Joseph. Estimating Betas from Nonsynchronous Data. Journal of Financial Economics 5, 1977, 309-327. Shaw, Wayne H. An Examination of Ex-Dividend Day Stock Price Movements: The Case of Nontaxable Master Limited Partnership Distributions. Journal of Finance 46(2), 1991, 755-776. Siddiqi, Mazhar. Ex-dividend returns and the Tax Reform Act of 1986. The Financial Review 32(1), 1997, 71-86. Skinner, David Lynn. Twenty-five years of tax law changes and investor response. Journal of Financial Research 16(1), 1993, 61-70. Stickel, Scott E. The Ex-Dividend Behavior of Nonconvertible Preferred Stock Returns and Trading Volume. Journal of Financial and Quantitative Analysis 26(1), 1991, 45-62. Stoll, Hans R. Inferring the Components of the Bid-Ask Spread: Theory and Empirical Tests. Journal of Finance 44, 1989, 115-134. Tinic, Seha M. and West, Richard R. Competition and the Pricing of Dealer Service in the Over-the-Counter Stock Market. Journal of Financial and Quantitative Analysis 7(3), 1972, 1707-1727. Walkling, Ralph A. and Edmister, Robert O. Are There Commission Cost Side-effects from Portfolio Management Decisions? Financial Analysts Journal July/August, 1983, 52-59.