CAPITAL GAINS TAXES AND IPO PRICING. Mary Ann Robinson* Richard M. Robinson* * Department of Accounting and Finance, Eastern Kentucky University

Transcription

1 Financial Decisions, Fall 2002, Article 1. Abstract CAPITAL GAINS TAXES AND IPO PRICING Mary Ann Robinson* Richard M. Robinson* * Department of Accounting and Finance, Eastern Kentucky University An hypothesis that personal tax rates on ordinary income and capital gains should influence the degree of IPO underpricing in predictable ways is presented. More specifically, it is shown that there should be an inverse relation between the capital gains tax rate and the degree of underpricing. An empirical analysis of the effect of the 1997 Taxpayer Relief Act, which lowered the capital gains tax rate, shows that the average degree of underpricing did increase as predicted, and that this occurs after controlling for other possible influences. Judiciously selecting the effective date of tax cuts allows empirical testing of competing popular underpricing-hypotheses without specification errors. Keywords: Initial Public Offereings; Capital Gains Taxes. JEL Classification: G1 I. Introduction The empirical evidence documenting IPO underpricing is voluminous. This evidence shows that disproportionate capital gains are typically earned on the first day of market trading. The gains are disproportionate in the sense that they exceed on average the amounts warranted by market risk. The existence of these gains implies money left on the table for the owner/entrepreneurs. One of the more popular hypotheses for explaining this phenomenon ascribes underpricing to underwriters fear that an issue might be under subscribed, i.e. underwriters price the issue below its anticipated market value to attempt to solicit an over subscription (at least a marginal over subscription). A struggle therefore exists between the underwriters and the owner/entrepreneurs with the former wanting below market price, and the latter desiring a higher price. The coercive strategy for each is the threat of withdrawn business, the owners threatening to withdraw the issue, and the investment bankers threatening not to underwrite. Of course the owners generally retain a portion of the firm so that they partially participate in the after-market capital gain resulting from the underpricing. Capital gains taxes affect the owners returns through both the actual IPO shares sold (where capital gains are measured as the difference between the IPO price and the original basis for the shares), and through the aftermarket selling of retained shares (here the capital gain is measured as the difference between 1

2 the after-market price and the original basis). 1 A cut in the capital gains tax rate should make underpricing more palatable for owners. It would make it easier for the underwriters to persuade the owners to accept an underpriced issue, and thereby lead to a higher degree of underpricing on average. The underlying motive for underpricing, however, remains an unresolved issue in the existing literature since any time a firm issues shares for less than they are worth, the new shareowners win and the original owners lose. Several authors [see Smith (1986), Tinic (1988), and Ibbotson, Sindelar and Ritter (1988), as example] present surveys of various hypotheses offered to explain the underpricing riddle. Tinic (1988) provides an in-depth explanation of the risk-averse underwriter, the monopsony power, the agency cost, the institutional lag, the speculative bubble, and the implicit insurance hypotheses. Johnson and Miller (1988) and Carter and Manaster (1990) explore the underwriter s prestige hypothesis. Reside, Robinson, Prakash and Dandapani (1994) pose a tax motive for underpricing. Hanley (1993), Benveniste and Spindt (1989), and Ritter (1987) pose a reward for information rationale for underpricing. Booth and Chua (1996) offer a liquidity rationale for underpricing. Among these competing hypotheses, the underwriter s prestige hypothesis has been extensively researched in recent years, and various indexes of prestige have been developed. [See Carter, Dark and Singh (1998) for a complete review and empirical investigation.] The general findings are that the greater the underwriter prestige, the lower the degree of underpricing. Beatty and Welch (1996), however, argue that this inverse relation recently reversed for small firms. Busaba, et al. (2001) confirm this reversal of underwriter-ranking influence. At first examination, a cut in capital gains taxes should equally rally both the IPO market and the secondary market. There appears no obvious reason why the degree of underpricing should be affected, i.e. a tax change should have a neutral effect on underpricing. Guenther and Willenborg (1999), however, show that the 1993 cut in capital gains rates for limited qualifying-companies resulted in a decrease in underpricing for the qualifying companies. But this tax cut (only 50 percent of the qualifying capital gain was taxable under this Act) only applied for shareholders who purchased new equity issued through IPOs, and then only for equity held for five years. This had the effect of rallying the IPO market, but not the aftermarket, and therefore underpricing decreased. Reside et al. (1994) is the only effort to date that actually offers an entrepreneurial tax motive for underpricing IPOs. They present an entrepreneurial wealth-maximizing model of tax motivated underpricing that is based on the deferred capital gains and associated taxes that occur through underpricing. They also present empirical evidence of this tax effect by measuring the impact of the 1986 Tax Reform Act on the degree of underpricing. Unlike the 1993 Act, the 1986 Act changed capital gains rates for both IPO and aftermarket purchases. As such, it is a more suitable event for examining the effects on underpricing. 1 The allowable aftermarket selling period for owners is frequently restricted to past 180 days after the IPO issue date. 2

3 The Reside et al. model allows owners to set the IPO price according to wealth maximizing criteria, i.e. the role of the underwriters in price setting is not recognized, and this a the weakness in their approach. An expanded model of IPO-generated owner-wealth, one that is somewhat similar to Reside et al (1994)., is presented in section 3. One significant difference in the model presented in 3 is that underwriters are allowed to establish the IPO price. This model utilizes a simulation that illustrates the potential effects of a capital gains rate cut on underpricing. 2 To be consistent with Reside et al., however, a presentation in the Appendix shows that if the owners do select an expected utility maximizing IPO price, then underpricing results. An empirical investigation of the impact of the 1997 Taxpayer Relief Act on IPO pricing follows in section 4. The tax-based influence on IPO underpricing is shown to exist and conforms to a priori prediction. The existence of the tax influence on IPO pricing is obviously relevant for empirical investigations of the competing underpricing hypotheses. As shown here, dummy variables that are judiciously selected to be consistent with the dates of the tax cut can measure the impact of the tax-rate changes in multivariate explorations that attempt to test the competing hypotheses. Omission of the tax-rate change leaves the multivariate analysis subject to misspecification errors. II. After-Tax Wealth From IPOs There are three potential sources of entrepreneurial wealth earned from an IPO: (1) capital gains earned from the distributed IPO shares sold to the public, (2) after-market capital gains earned from owner-retained shares, and (3) any owner-distributed surplus earned from the issue. Each is taxed differently: 3 The capital gain constituted by the difference between the after-market price and the original basis paid by the owners on the retained shares has a tax liability that is deferred until the sale is realized, and the applicable tax rate is the capital gains rate. The capital gain constituted by the difference between the IPO issue price and the original basis paid by the owners on the shares sold to the public is also taxed at the capital gains tax rate. Any anticipated surplus above the cost of projects that might be funded by the proceeds from the IPO, and that is distributed to the owners in the form of a pre-ipo dividend, is taxable at the ordinary income tax rate. Good examples of surplus distributions are provided by the 1998 IPOs of Columbia Sportsware Co., Contacts Inc., and Miller Exploration (details of each are provided here). 2 The model is expanded as compared to Reside et al. (1994) in that current (IPO time) capital gains are included here, and both current and after-market capital gains are measured against the properly formed basis according to tax law. 3 The capital gains rates referred to below will apply provided that the investment s holding period is of the appropriate length as indicated in section 3. 3

4 Examples of the surplus distribution are abundant in recent IPO history. Some are worth citing here because this method of IPO wealth generation is not generally recognized in the literature. Columbia Sportsware Co. was a Subchapter S corporation prior to its IPO on March 27, Prior to the IPO issuance, the company declared a dividend equal to the greater of $95 million or the amount of its Subchapter S accumulated adjustments account. The dividend was payable to the pre-ipo existing S-Corporation shareholders. The net proceeds from the IPO were used to pay the dividend. (See Columbia s S-1 Report for 1998.) A dividend distribution similar to Columbia s was made by Contacts Inc. Their dividend to S-Corporation shareholders, however, was distributed in the form of a promissory note. The firm went public on February 10, 1998, and proceeds from the IPO distribution were used to fully redeem the promissory notes before the end of that year s first quarter. Miller Exploration went public on February 4, Prior to the IPO, it declared bonuses to firm managers (who collectively owned 30 percent of the pre-ipo shares) to be paid from the IPO proceeds. Of the three indicated sources of owner wealth (the three bulleted above), the latter two indicate that owner wealth is enhanced by a higher IPO price, but the first indicates that owner wealth is enhanced by a lower IPO price. Underwriters, of course, determine the IPO offer price after soliciting potential interest from institutional investors through their roadshow or bookbuilding activities. One suspects that the owner/entrepreneurs probably prefer a higher IPO price since this enables them to exploit higher current capital-gains and/or surplus income, the bird-in-the-hand motivation. A lower current price, however, can be recouped in the after market through a deferred capital gain. For this reason, owners resistance to underpricing may be relatively weak in that it takes a considerable expected degree of underpricing before they withdraw the issue. 4 [See Busaba, et al. (2001) for an analysis of factors that determine the probability that a prospective IPO is withdrawn.] The underwriter is motivated to underpice in order to easily obtain a fully subscribed issue, or to reward investors for information production during the bookbuilding period. The lower the capital gains tax rate, the more attractive the aftermarket capital-gain becomes to the owners. For this reason, the underwriters may be able to force a higher degree of underpricing after a cut in capital-gains tax-rates without the owners withdrawing the issue. The simulation presented in the next section shows that owner wealth can increase when IPO price and capital-gains tax-rates are simultaneously cut. This illustrates the point that underwriters may be able to force this higher degree of underpricing when rates are cut. This provides a basis for an empirically testable hypothesis since the Taxpayer Relief Act of 1997 (TRA) reduced capital-gains tax rates. 4 Busaba et al. (2001, p. 76) quote the president of Kwik Goal concerning their withdrawn IPO. They withdrew because the final IPO offer price was overly low as compared to the owner s expectation of market value. 4

5 III. The Taxpayer Relief Act of 1997 and Simulation A. Applicable Tax Rates on Capital Gains Prior to August, 1997, capital gains were taxed at the taxpayer s ordinary-income tax rate up to a maximum of 28 percent as specified by the Tax Reform Act of The Taxpayer Relief Act of 1997 (TRA), however, reduced the capital gains tax rate to as low as 10% in some circumstances. The TRA defines three possible holding periods for assets, and also specifies the tax rate(s) that apply to each: (1) short-term (one year or less) capital gains are taxed at the taxpayer s ordinary-income tax rate; (2) mid-term (more than one year but less than or equal to eighteen months) capital gains are taxed at either a 28% rate or the taxpayer s ordinary-income tax rate, whichever is lower, and (3) long-term (more than 18 months) capital gains are taxed at 10%, 20%, or 25% depending on (a) the type of asset, and (b) the taxpayer s marginal tax bracket. [See Hoffman, Smith and Willis (1999) for a detailed description of the application of these tax brackets.] These tax rates apply to capital assets sold after July 28, 1997, but before January 1, The Taxpayer Relief Act also set the maximum tax rate for mid-term capital gains to be 20% after January 1, The rate was specified to change again for capital assets purchased after 2000 and held for at least five years. The TRA requires a somewhat complex process for determining the appropriate capital gains tax rate. This process can yield up to fourteen scenarios to which apply a combination of different rates. Consider, however, two simple outcomes of the process for assets sold after July 28, 1997, but before January 1, The two outcomes are determined by classification as either longterm or mid-term capital gains. As illustrated by Table 1, mid-term capital-gain rates are less than ordinary-income tax rates in four of the five tax brackets, and long-term capital gain rates are below ordinary-income tax rates for every bracket. The tax rate on long-term net capital gains is almost one-half that of ordinary income for taxpayers in the highest tax bracket. Note that for assets sold after December 31, 1998, all mid-term gains realized by individuals in tax brackets above 15% have a capital gains tax rate of 20% instead of 28%. While the Taxpayer Relief Act is certainly complex concerning applicable capital-gains tax rates, Table 1 clearly shows that in almost all classifications, rates were reduced. IPO owners could certainly expect to exploit these classifications. The after market was expected to rally as a result of the tax cut, and this should have affected IPO prices as well. B. Simulation of Underpricing and Tax-Rate Change It is argued above that the Taxpayer Relief Act of 1997 should have lowered owner resistance to the underpricing of IPOs as forced by underwriters. Since underwriters are motivated to underprice, this should result in a higher average degree of underpricing. To simulate the effects of changes in both the tax rate and underpricing amount, allow W to be the owner wealth earned from the IPO. Also allow the additional following symbology: 5

6 Table 1: Capital Gains Tax Rates as Dependent on Ordinary Income Tax Rates T o is ordinary income tax rate. T c is capital gains tax rate. Ordinary Mid-term Long-term Income net capital gains* net capital gains T o T c T c 15% 15% 10% *After 1998, all mid-term tax rates are 20% for ordinary-income tax brackets above 15%. V = uncertain (random) aftermarket total value of shares, C = cost of any firm project financed with proceeds from IPO (these projects are frequently listed in the IPO prospectus), P = total value of IPO issue, B = original basis owners paid for shares as defined by tax law, α = proportion of firm retained by the entrepreneur (0 α 1), T o = entrepreneur s personal tax rate on ordinary income, T c = entrepreneur s tax rate on capital gains income (T c T o ), γ = a present value interest factor to measure the present value of deferred tax on realized capital gains (0 γ 1), over the entrepreneur s desired deferral period, where, if γ = 1, capital gains are not deferred and are realized in the current period, and if γ = 0, the capital gains are deferred indefinitely and the capital gains taxes are completely avoided. The projects referred to above (with costs equal to C) may need some explanation. When IPO shares are issued, the prospectus frequently mentions some project that will be finance from the proceeds. The project may be plant expansion, or existing debt retirement, or inventory purchase, or some other asset purchase. The amount raised by the sale of shares is typically in excess of that needed to finance the project, and the proceeds are usually used as a dividend to the owners (distributed surplus), as reviewed in section II. Owner wealth generated by the IPO comes from three sources. The first is the distributed surplus of (P C), taxed at the ordinary income tax rate: (P C)(1 T o ). The second is the capital gain 6

7 earned from the IPO sale of owners shares, taxed at the capital gains tax rate: (P B)(1 - α)(1 T c ). The third is the owners retained share of the aftermarket capital gain taxed at the capital-gains tax-rate but with the tax deferred until realized: (V B)α(1 γt c ). At the time of the IPO issue, this owner-wealth increment is uncertain since V is uncertain. As a result, the future tax liability is also uncertain. With the deferral present-value factor being given by γ, at the IPO issue time this uncertain tax liability is (V B)α γt c. The owner wealth generated by the IPO is given by (1). W = (P C)(1 T o ) + (P B)(1 - α )(1 T c ) + (V B)α(1 γt c ) (1) Equations (2) and (3) present the first and second derivatives of IPO price with respect to capital-gains tax-rate along an isowealth function where wealth is held to W 0 while T c and P are allowed to vary. (For this initial analysis, V is not allowed to vary with T c. This is changed below.) The isowealth function is depicted by Figure 1. As shown, a decrease in T c can be offset by a decrease in P, one that is consistent with (2), so that W is unchanged. This tax cut is illustrated by Figure 1 where T c * is an initial tax rate, and T c ** is the rate after a cut. P* is the initial IPO price, and P** is the price that holds wealth at W 0 after the tax cut. Any associated IPO price reduction that is less than P* - P** increases wealth. The implication is that following a tax cut, underwriters could reduce IPO prices by some amount, thereby increasing the degree of underpricing, while still allowing an increase in owner wealth. P T c (P B)(1 α) + γ(v0 B)α = > 0 (2) 1 T0 + (1 α)(1 T c ) W=W 0, V=V 0 2 P T c 2 ( P/ T )(1 α) = c > 0 (3) 1 T0 + (1 α)(1 T c ) W=W 0, V=V 0 A reduction in capital-gains tax-rates should, of course, rally the aftermarket. Allowing V to adjust to the tax rate, the derivative of the isowealth function is given by (4), where it is P P assumed that MV/MT c < 0 so that <. T c T W=W 0, V=V 0 c W=W 0, V > V 0 P T c (P B)(1 α) + γ(v0 B)α - α(1- γt ) V/ T = c c (4) W=W 0, V=V 0 1 T0 + (1 α)(1 T c ) > 0 where MV/MT c < 0 7

8 If we now allow the isowealth function depicted by Figure 1 to apply where the after market is allowed to rally, MV/MT c < 0, then it is apparent that a reduction in the tax rate from T c * to T c ** where there is no reduction in IPO price, not only increases wealth, it increases the degree of underpricing. Figure 1: An isowealth function is depicted where W is held to W 0 while P and T c change according to (2). Lowering T c from T c * to T c **, and P from P* to P** leaves wealth at W 0. P W=W 0 P* P** B P γ(v B) α = > 0 T c 1 T P=B 0 T c ** T c * T c For the function depicted, P is restricted to P B. To simulate (illustrate) the effects of a simultaneous cut in tax rates and IPO price, but while holding V constant, allow the following initial values for the above variables: C = $50, B = $40, V = $110, " = ½, ( =.6209 (this value of ( corresponds to a defral period of 5 years and an APR required rate of return of 10%). Further initially assume that T 0 = T c = 28%, and that the IPO price is P = $100. With this simulation and using equation (1), W = $ For the second simulation, decrease the capital gains rate to T c = 20%, and the IPO price simultaneously to P = $98. For this simulation, wealth increases to W = $ Indeed, any price above increases owner wealth. 8

9 Comparing these simulation results illustrates that a simultaneous cut in both capital-gains tax rates and IPO price can increase wealth. Allowing an increase in V to accompany the cut in T c makes the underpricing proposition stronger. The implication is that when the tax rate is reduced, underwriters may be able to increase underpricing without the owners withdrawing the issue. This provides the impetus for the empirical test presented below. A. Sample Data IV. Empirical Methodology Empirical analyses of IPO underpricing generally study the initial excess returns, i.e. returns in excess of the market average rate of return. Particularly, the IPO return is measured from offer price to either the secondary market s first-trade price or first-closing price, and this return is adjusted by the return on some market index. [See Carter, Dark and Singh (1998) for an example of the use of this market index in IPO research.] Although this method does not explicitly consider the market-related risk associated with the individual stock issue, it does allow adjustments for market effects on the returns of IPOs. [See McDonald and Fisher (1972), Beatty and Ritter (1986), and Ritter (1991) for early usage of this method.] Empirical investigations of IPO returns have included underwriter rankings, asset size, age of the firm, debt-to-asset ratio, per share offer price, the dollar amount of the issue, and the number of purposes listed for the IPO proceeds as just some of the independent variables in regressionbased multivariate analyses. For example, Tinic (1988) included the reciprocal of the offering value, the natural log of the offering value, and dummy variables for ranked versus nonranked underwriters and for issuing in the month of January. The dependent variable was marketadjusted returns. Only the coefficient for the reciprocal of the offering price was significantly different from zero. Beatty and Ritter (1986) included the reciprocal of the offering price, and an underwriter prestige index as independent variables, along with the stated number of uses for the funds raised as a proxy for uncertainty concerning the returns to the investor. This latter variable was found to have no effect. The coefficient for the reciprocal of the offer price was significantly different from zero. The dependent variable was the market adjusted rate of return. Reside et al. (1994) included underwriter prestige ranking, total assets, debt-to-assets ratio, the age of the firm, the value of the issue, and a dummy variable for the capital gains tax change of All variables were found to have coefficients significantly different from zero except total assets and the value of the issue. Beatty and Welch (1996) included underwriter ranking, underwriter compensation, the number of risks stated in the prospectus, the number of stated uses for the proceeds, insider retention, and the inverse of the offer price as explanatory variables. All coefficients were significantly different from zero except insider retention and the number of uses. 9

10 Brennan and Hughes (1991) included the reciprocal of the per share offer price as one independent variable. They argue that lower share-price increases brokerage commissions, and may also be an index of quality. An obvious problem with this measure, however, is that it may itself be a simple index of underpricing, and its use as an explanatory variable could be questioned. Two different underwriter ranking indexes have generally been used in IPO studies: the prestige index of Carter and Manaster (1990), and the market share index of Megginson and Weiss (1991). The Carter-Manaster ranking comes from the underwriter s relative placement in tombstone advertisements in The Wall Street Journal. Larger underwriters are generally listed at the top of the advertisement, and in larger type. Small underwriters are listed at the bottom in smaller type. The Megginson-Weiss index is a simple measure of market share. Carter, Dark and Singh (1998) provides an in-depth explanation and empirical exploration of these indexes. The regressions reported below use the initial market-adjusted rate of return, as measured over one-day and five-day periods, as the dependent variables. The independent variables include the reciprocal of the offer price, the value of the firm's assets, the log of assets, the underwriter's ranking as indexed by both Carter and Manaster (1990) and also Megginson and Weiss (1991) [using the updated rankings for both as provided by Carter, Dark and Singh (1998)], and a dummy variable of "1" for issues after July, 30, 1997, and "0" for those issued before. This event date was chosen since bills involving the decrease in the capital gains taxes passed both Houses of Congress, and the Joint Tax Committee of Congress on this date. The compromise bill became law shortly after on August 5, 1997, but the president already announced his intention to sign when the bills were passed on July 30. [See O Connell (7/30/97) and Wessel (7/31/97).] 5 All NASDAQ listed IPOs issued from January 1, 1995, through June 30, 1999, were included in the data studied provided that the issues were of initial, stock-only offerings. As Ritter (1984) and Beatty and Welch (1996) hypothesized, size may influence the risk of the issue. NASDAQ issues vary considerably in firm size, but NYSE IPOs are restricted to large firms. Restricting the IPOs studied to stock-only offerings is important. Common stocks issued in unit offerings or mixed debt-equity offerings are omitted from the sample since the return process for them may reflect the behavior of the other financial instruments in the offering, and not of the common stock itself. Consequently, the data consists of NASDAQ common stocks that were offered without warrants and without mixed debt-equity packages.the resulting data sample examined consists of 605 IPOs of firms added to the list in Standard & Poor s NASDAQ Daily Stock Quotations, 444 IPOs prior to the July 30, 1997, event date, and 161 IPOs after that date. 5 Other dates explored included the announcements of backing by various key congressional members (mid-january and Mid-February) and the published descriptions of various tax plan details (end of February and beginning of May and July). These alternative dates did not yield significant findings. 10

11 Closing prices on the first day of trading were gathered from the same source, as were values for the NASDAQ Industrial Index. IPO offering prices were gathered from Moody s OTC Industrial Manuals. Firm characteristics such as the date of incorporation, total assets, underwriter, and the date of the offer were gathered from the same Moody s source. Table 2 reports the mean-excess returns for the sample of IPOs both prior to and after the July 30, 1997 event date. As shown, t-statistics (t*) indicate that both one-day and five-day excess returns are significantly above zero for each period. A t-test (t**) also indicates that both oneday and five-day mean excess returns increased after the July 30, 1997 date. Table 2: Mean Excess One-Day and Five-Day Returns Before and After July 30, Pre June 30, 1997 Post June 30, 1997 One Day Returns 22.22% 21.18% t* Five Day Returns 27.57% 24.81% t* t** * t-statistics measure differences of means from zero. ** t-statistics measure differences between Pre and Post June 30, 1997 means. B. Multivariate Examination of Underpricing Table 3 presents descriptive statistics for the variables included in the regressions reported here. In addition to the listed variables, debt ratios, real values of assets, and the age of the firm at IPO date were also included in regressions, but these variables were found to have no impact on underpricing. They are not included in the reported regression here. It is assumed that in the aftermarket, the issue price adjusts quickly from the IPO price to the fair market valuation. For this reason, one-day excess returns and five-day excess returns were measured (in excess of the 11

12 Table 3: Descriptive Statistics One-Day XR is the market adjusted one-day excess return. Five-Day XR is the five day market adjusted excess return. CM Rank is the Carter-Manaster underwriter ranking. MW Rank is the Megginson-Weiss underwriter ranking. Assets are book values in $1,000. 1/Off is the inverse of the per-share offer price. Variable Mean Median Std. Deviation One-Day XR Five-Day XR CM Rank MW Rank Assets $42,722 $20,480 $73,905 1/Off NASDAQ Industrial Index). For the purpose of examining possible multicollinearity problems with the independent variables, Table 4 presents the correlation coefficient matrix of the variables used in the regression. Note that, although the coefficients between the CM ranking and the MW ranking is high, these variables were included separately in the computed regressions. Tables 5 and 6 show the results of OLS regressions using standardized one-day and five-day returns. Regressions were also computed using nonstandardized returns, but the hypotheses of normality for the residual errors were rejected at 5% confidence levels by χ 2 measures. Normality could not be rejected for the standardized excess returns. 6 Tables 5 and 6 do not show the results for all of the independent variables tried. The age of the companies (time since incorporation), the real values of assets, and debt/asset ratios were also tried. The coefficients for the age, value of assets and debt/asset ratios were not significantly different from zero at even the 10% level, and did not improve the explanatory power of the regressions as measured by F ratios; nor did they alter the magnitudes or significance levels of 6 Given the central-limit theorem, standardizing the dependent variable improves the possibility of obtaining normally-distributed residuals. 12

13 Table 4: Correlation Coefficients 1-Day is the market adjusted one-day excess return. 5-Day is the market adjusted five-day excess return. CM is the Carter-Manaster underwriter ranking. MW is the Megginson-Weiss underwriter ranking. Asset is the book value of assets. 1/Off is the reciprocal of the per share offer price. 1-Day 5-Day CM MW Asset 1/Off 1-Day Day CM MW Asset /Off 1.00 other variables. Also, deflating asset values did not materially contribute to the regression results. Since heteroscedastic disturbances could obviously be associated with the event date, the Goldfeld-Quandt test was computed for residual variance comparisons for IPOs issued before and after July 31, The hypothesis of homoscedasticity could not be rejected at 5% significance levels. The OLS results shown in Tables 5 and 6 indicate an inverse relationship between the magnitude of assets and the degree of underpricing. The larger the company s assets, the lower the one-day and five-day excess returns, i.e. the lower the degree of underpricing. These results are generally consistent with other studies. The coefficients for the inverse of the offer price are also negative, and significantly different from zero. This is consistent with Beatty and Welch s (1996) findings. Also consistent with Beatty and Welch (1996), and Busaba, et al. (2001) are the positive underwriter-ranking coefficients. This contradicts previous studies that use older data. Beatty and Welch show that these positive coefficients are associated with smaller firms, and argue that this is a development of the 90s. 7 Given our NASDAQ sample, this small firm argument is consistent with the findings reported here. 7 After controlling for size using a principal components analysis, Beatty and Welch (1996) find a positive relationship between underpricing and underwriter ranking, and a negative relationship between underpricing and the inverse of the offer price for small firms for the period

14 Table 5: OLS With One-Day Standardized Excess Returns Constant (6.29) (2.47) (6.16) (4.47) (0.61) (2.82) MW Rank (2.18) (1.84) (3.61) CM Rank (0.98) (0.94) (3.14) Assets* (4.71) (4.57) (3.66) (3.56) Log Assets* (4.30) (4.39) 1/Off (7.05) (7.00) (6.82) (6.75) Dummy (1.96) (1.82) (2.10) (2.03) (2.18) (2.03) R % 11.8% 11.8% 11.4% 4.5% 4.0% DW F** t-statistics are in parentheses. 1 Significant at 1%. 2 Significant at 2.5%. 3 Significant at 5%. * Assets are measured in $ millions ** F-ratio is for inclusion of dummy variable. Given that the 1/Off variable and CM and MW rankings have some degree of negative correlation, the regression results without the 1/Off variable are shown. The signs of the ranking coefficients do not change with the omission of 1/Off. For all specifications tested, the dummy variables for the tax date have positive coefficients as the tax model predicts, and all coefficients are significantly different from zero at 5% levels at least. Also, the F-ratio for inclusion of the tax date dummy (included with MW Rank, Assets, and 1/Off. Pr) is 3.73 for for the one-day and five-day excess return regressions. Both are significant at the 1% level. The hypothesis that the cut in capital gains taxes increases one and five-day excess returns cannot be rejected at any reasonable level of significance. 14

15 Table 6: OLS With Five-Day Standardized Excess Returns Constant (6.54) (2.59) (6.27) (4.55) (0.02) (2.67) MW Rank (2.18) (1.84) (3.61) CM Rank (0.98) (0.94) (3.14) Assets* (4.71) (4.57) (3.66) (3.56) Log Assets* (4.30) (4.29) 1/Off (7.05) (7.00) (6.82) (6.75) Dummy (1.98) (1.86) (2.10) (2.03) (2.18) (2.03) R % 11.8% 11.8% 11.4% 4.5% 4.0% DW F** t-statistics are in parentheses. 1 Significant at 1%. 2 Significant at 2.5%. 3 Significant at 5%. *Assets measured in $ millions ** F-ration is for inclusion of dummy variable. V. Discussion and Conclusion The tax influence on the degree of underpricing does not refute other underpricing hypotheses such as the risk-based hypothesis mentioned above. It should be viewed as an additional influence on the degree of underpricing. By underpricing the issue, the underwriter forces the entrepreneur to accept lower current income from the IPO in return for deferred capital gains. The current income is immediately taxable, but the aftermarket capital gain is tax deferred until realized. Preferential capital-gains tax-rates should lower the entrepreneur s resistance to trading current income for deferred capital-gains that result from underpricing. The 1997 reduction in capital-gains tax-rates offers an event-date methodology for testing the tax-based influences. The tax-based argument presented above hypothesizes that a cut in capital gains rates should cause an increase in underpricing. After controlling for other explanatory variables 15

16 included in other IPO studies, the tax-based hypothesis is empirically supported, i.e. the capitalgains tax-cut of 1997 did increase the average degree of IPO underpricing. This tax effect is obviously relevant for empirical investigations of the various IPO underpricing phenomenon. Omission of controls for tax rate changes exposes any multivariate study to misspecification errors. Appendix A: A Tax-Based IPO Model Like the Reside, et al (1994) model, the model presented here has the owner/entrepreneur setting the price of the IPO. Under conditions of uncertainty with respect to the after-market price, it is shown that underpricing the issue maximizes owner s expected utility. Allow the use of the same symbology as in section 3. Further allow U(W) to be the utility of wealth where U conforms to the von Neumann Morgenstern axioms for expected utility, with MU/MW = U > 0, and U < 0. Also allow p(v) to be the probability density function for the random after-market price. To obtain an interior solution for P, also assume that α = f(p), with f < 0, and f > 0. This functional relation occurs because owners would be expected to lower their retained portion as price rises due of a desire to exploit the higher current capital-gain income by selling their shares. The higher price means lower deferred capital-gains to be exploited by the retained portion. The owner wealth generated by the IPO is still given by equation (1). Expected utility is given by (2). Equation (4) establishes a first order condition for maximizing expected utility. W = (P C)(1 T o ) + (P B)(1 - α )(1 T c ) + (V B)α(1 γt c ) (1) + EU(W) = U(W)p(V)dV (2) ME(U)/MP = E(U ){(1 T o ) + (1-α )(1 T c ) + f [(EV B)(1 γt c ) (P B)(1 T c )]} + + f Cov(U, V)(1 γt c ) = 0 if (3) X P + B + (EV B)Y = - Cov(U, V)/E(U ) (4) where X = [(1 T o )/ (1 T c ) α]/ f > 0 Y = (1 γ T c )/ (1 T c ) A simplification of (4) readily allows an analysis of the degree of underpricing, one that presents various insights into the undepricing phenomenon. Assume that capital gains taxes on the 16

17 aftermarket wealth are not deferred, that they are due immediately at the IPO issue date even if they are unrealized in the sense of having the retained shares sold, i.e. γ = 1. Under this condition, Y = 1, and equation (4) reduces to (5) where the degree of expected underpricing is shown to be positive. This follows since f < 0, and also Cov(U, V) < 0 because of diminishing marginal utility to wealth. EV P = - X - Cov(U, V)/E(U ) (5) > 0 It is tempting to analyze the effects of a cut in T c on EV P, but a cut in taxes changes wealth, and therefore U, as well as X. The results are ambiguous. The weakness in this model s approach is apparent considering that the IPO owners do not have full discretion over the setting of the issue price. This is also the weakness in Reside et al s (1994) effort. Most of this pricing discretion lies with the underwriters, while the owners do influence the price by their underwriter selection ability, and their ability to withdraw the issue. 17

18 References Beatty, R.P. and J.R. Ritter (1986), Investment Banking, Reputation, And The Underpricing Of Initial Public Offerings, Journal of Financial Economics 15, pp Beatty, R.P. and I. Welch (1996), Issuers Expenses and Legal Liability in Initial Public Offerings, Journal of Law and Economics 39, pp Benvenite, L. and R. Spindt (1989), How Investment Bankers Determine Offer Prices, and Allocation of New Issues, Journal of Financial Economics 24, pp Booth, J. and L. Chua (1996), Ownership Dispersion, Costly Information, and IPO Underpricing, Journal of Financial Economics 15, pp Brav, Alon and Paul A. Gompers (2000), Insider Trading Subsequent to Initial Public Offerings: Evidence from Expirations of Lock-Up Provisions. Working Paper. Bradley, Daniel J., Bradford D. Jordan, Ivan C. Roten, and Ha-Chin Yi (2000), Venture Capital and IPO Lockup Expiration: An Empirical Analysis Journal of Financial Research XXIV, No. 4. Brennan, M.J. and P. Hughes (1991), Stock Prices and the Supply of Information, Journal of Finance 46, pp Busaba, W., L. Benveniste, and R, Guo (2001), The Option to Withdraw IPOs During the Premarket: Empirical Analysis, Journal of Financial Economics 60, pp Carter, R.B., F. H. Dark and A.K. Singh (1998), Underwriter Reputation, Initial Returns, and the Long-Run Performance of IPO Stocks, Journal of Finance 53, pp Carter, R.B., and S. Manaster (1990), Initial Public Offerings and Underwriter Reputation, Journal of Finance 45, pp Guenther, D. and M. Willenborg (1999), Capital Gains Tax Rates and the Cost of Capital for Small Business: Evidence From the IPO Market Journal of Financial Economics 53, pp Guo, R. (1998), On the Cost of Withdrawn Initial Public Offerings. Working Paper, University of Minnesota. Hanley, K. (1993), The Underpricing of Initial Public Offerings and the Partial Adjustment Phenomenon, Journal of Financial Economics 34,

19 Hoffman, W.H., J.E. Smith and E. Willis (1999), Individual Income Taxes. (West/South-Western College Publishing, Cincinnati). Ibbotson, R., J.L. Sindelar and J.R. Ritter (1988), Initial Public Offerings, Journal of Applied Corporate Finance 1, pp Johnson, J. and R. Miller (1988), Investment Banker Prestige and the Underpricing of Initial Public Offerings, Financial Management 17, pp McDonald, J.G. and A.K. Fisher (1972), New-Issue Stock Price Behavior, Journal of Finance 27, pp Megginson, W. and K. Weiss (1991), Venture Capitalist Certification in Initial Public Offerings, Journal of Finance 46, pp O Connell, V. (1997), What the Tax Agreement Means to You: How to Save Most from New Rates, The Wall Street Journal (July 30): C1. Reside, M.A., R.M. Robinson, A.J. Prakash, and K. Dandapani (1994), A Tax-Based Motive for the Underpricing of Initial Public Offerings, Managerial and Decision Economics 15, pp Ritter, J.R. (1991), The Long-Run Performance of Initial Public Offerings, Journal of Finance 46, pp Ritter, J.R. (1984), The Hot Issue Market of 1980, Journal of Business 57, pp Smith, C.W. (1986), Investment Banking and the Capital Acquisition Process, Journal of Financial Economics 15, pp Tinic, S.M. (1988), Anatomy of Initial Public Offerings of Common Stock, Journal of Finance 43, pp Wessel, D. (1997), Gains-Cut Plan Is Modified by the GOP, The Wall Street Journal (July 31): A2. 19