This paper was published in the Journal of Management and Business Review, PPM School of Management

Transcription

1 UNDERPRICING, MARKET SHARE, AND EX ANTE UNCERTAINTY OF UNDERWRITING Bramantyo Djohanputro, PhD Lecturer and Consultant of Finance, Investment, and Risk Management This paper was published in the Journal of Management and Business Review, PPM School of Management Absctract This study focuses on three issies. The first concerns with the pricing behaviour of new issues. This is to evaluate the level of underpricing and the pricing across underwriters. The second is the study of the relationship between underpricing and market share. And the third is the relationship between underpricing and ex ante uncertainty. This study indicates that the price movement follows the efficient market hypothesis immediately after the stocks being traded in the market. It is also found that there are two proxies that are significantly related to ex ante uncertainty. Those are the reciprocals of both book value and proceeds. The fact that nonprestigious underwriters cannot outperform prestigious underwriters in terms of the number of underwriting demonstrate that the deviations alone cannot explain the change in market share INTRODUCTION As it is well known, initial public offerings (IPOs) are on average underpriced, mostly based upon shortterm returns. However, the underpricing levels do not have a significant predictive value for investors to make decisions about the market price with because the price movement follows efficient market rules immediately in the after market (McDonald and Fisher, 97). If the returns are measured from the initial price to the firstweek trading price or the onemonthafteroffering trading price, the initial prices indicate a significant appreciation. But there is a very shortterm adjustment within one week of public trading so the appreciation from one week to one month is not significant (Neuberger and Hammond, 97). This implies that investors buying stocks in the initial offering obtain excess profit whereas investors cannot gain excess returns in the after market (Miller and Reily, 987). It is also found that the pricing behaviour among underwriters is different. Underwriters are classified into three main groups: special bracket, major bracket and submajor bracket (Hayes, 97). By using special bracket firms as prestigious underwriters and submajor

2 bracket firms as nonprestigious underwriters, it is concluded that offerings underwritten by prestigious underwriters are consistently less underpriced than those underwritten by nonprestigious ones. The consistency, unfortunately, does not necessarily confirm the relationship between the prestige and underpricing levels since the prestige level does not have explanatory power in its own right (Johnson & Miller, 988). Instead, the difference in underpricing levels is caused by the difference in the level of ex ante risk. There is an equilibrium price at which every issue satisfies all parties involved in underwriting. The equilibrium price is the price at which the issues offered are fully subscribed. The parties under consideration of equilibrium are underwriters, issuers and investors. The importance of considering issuers and investors in setting the prices is based on the need for underwriters to maintain their relationship with the issuers, investors and potential issuers. Suppose the price is too high, then it will be favourable to issuers because they can obtain huge proceeds. However, it is against the investors interests to maximize or gain excess returns at their expenses. Consequently, the investors will leave the underwriters alone in the future. Suppose, on the other hand, the price is too low. It will be favourable to investors seeking high excess returns, but it will be against issuers seeking to maximize their proceeds. Accordingly, the issuers are likely to leave the underwriters. Also, after knowing about the low performance of the underwriters, the potential issuers are unlikely to use the underwriters service. Thus, setting a right price must be related to future performance in terms of market share of underwriting business, and of the perception by both issuers and investors markets. The main powerful variable explaining the underpricing level or equilibrium is ex ante uncertainty about the value of the issue perceived by investors. The greater the ex ante uncertainty of a firm, the larger is the expected underpricing (Beatty and Ritter, 98). To indicate the level of uncertainty, The use of the number of uses as the proxy is based on the regulations in the USA, i.e. SEC requires, issuers to provide information about what the proceeds will be used. Firms under the classification of speculative firms by SEC are required to provide relatively detailed information about the use of proceeds. This requirement does not apply to established firms. This requirement can be employed as the proxy of the ex ante uncertainty. Highrisk firms with large ex ante uncertainty tend to have a list with the greater number of uses than lowrisk firms.

3 An underwriter who successfully develop its reputation have a great chance to be employed by issuers for the repetitive offerings. The repititive offering is defined as the use of same underwriter by an isuers for its seasoned offering. The repetitive offering follows four hypotheses. Firstly, the likelihood of subsequent offering is negatively related to IPO risk. This means that high quality firms are more likely to offer seasoned stocks than lowquality firms. Second, the likelihood of subsequent offering is positively related to the IPO underwriters reputation. This implies that if underwriters perform well in IPO, they will be employed again by the same firms to issue seasoned stocks. Third, the likelihood of subsequent offering is negatively related to IPO gross spread, that is the difference between the price agreed with issuers and the price offered to investors by underwriters. And finally, the likelihood of firms changing the underwriters in seasoned offering is negatively related to the IPO underwriter reputation. These assumptions have been proved by Carter (99). In order to offer seasoned or repeated issues, there are some factors to be considered. First, only highquality firms indicated by low ex ante uncertainty are willing to offer subsequent issues. To offer the subsequent issues successfully, then, the IPO s must be underpriced so the repeated offers can be priced more highly (Welch, 989). Low quality firm, on the other hand, tend to maximize their proceeds from IPOs that are less underpriced. Second, from the underwriter s point of view, the transaction costs in terms of the spread charged from the IPOs are very important for the continuity of the contract. Charging high transactions would discourage the firms from using the same underwriters because of high competition (Carter, 99, James, 99). Hence, it is argued that if the firms are likely to market the subsequent issues, the underwriters need to charge the costs for the IPO under the marginal costs and then increase the costs of the following offerings. Lastly, there is a need for underwriters to maintain their relationship with highquality firms to assure the future prestige that is able to create a self fulfilling effect. This means that while high quality firms tend to offer their issues employing prestigious underwriters, the underwriters use these contracts to build up their own reputation. As a result, other lowrisk firms will be attracted to make contract with these underwriters.

4 The analyses above show that the most important factors in evaluating the underwriting performance are the prestige of the underwriters, the level of underpricing, and the ability to predict the level of ex ante uncertainty. These are the interests of the following analysis. METHODOLOGY Classification of underwriters Classification of the underwriters in this study is based on Hayes method. Basically, the hierarchy is divided into tree brackets: special bracket, major bracket and sub majors and others. In certain underwriting syndicates, there may be ' major out of order' firms under the heading of major bracket. They are treated as major bracket firms in the syndicate but separated from them in the list because they are not be accepted as permanent major bracket members by the manager. Usually, major out of order firms are recommended by the issuers. The classification of underwriters can be identified based on two main indicators. The first indicator is the proportion of underwriting participation as the ultimate result of underwriting. In any offering in the US, for example, a special bracket participates the largest portion of the value of offering. Any major bracket firm participates around 0 per cent less than the amount of participation by the special bracket firm. Sub major bracket firm participates around to 70 per cent of a major bracket firms participation. As an alternative of using the proportion of participation, the ultimate result can also be identified through the alphabetical order of the names used in the advertisement for underwriting. The first underwriter in the list is the manager. It is followed by comanagers if any. Following the manager and/or comanagers are special bracket firms that are arranged alphabetically. They are followed by major bracket firms that are arranged alphabetically also, followed by the lost of major out of order firms and, at the bottom of thje list, sub major and other firms. It should be noted that there is a bias in such classification caused by the preferences of the underwriting manager or issuer. Usually, the manager will choose and suggest certain underwriters to be the members, while issuers prefer certain underwriters to be members of the syndicate and to participate for a certain proportion of the

5 underwriting. The issuer s preference induces an ad hoc participation in a syndicate under a certain manager. To adjust the bias, the analysis shall apply multiadvertisements evaluation. The ads chosen are only those that involve three brackets placed in alphabetical order. Then the positions of each underwriter are compiled to give a pattern of perceived prestige. The second indicator is the capabilities of the underwriters that relate to the ability to negotiate with clients. This indicator can be represented by the amount of branch offices as the proxy. Note that this proxy is logically not strong. In fact, the number of branch offices has a weak relationship with the position in the hierarchy. Besides, most special bracket firms do not have branch offices. Even if they have branches, the number is much less than the number of branches owned by other underwriters. For further analysis, the data about underwriters are classified into two categories, prestigious and nonprestigious underwriters. The former consists of special and major bracket firms while the latter consists of sub major bracket and other firms. The selected nonprestigious underwriters are only those who underwrite one equity offering in a year. Both categories amount to 8 underwriting firms. The level of underpricing The excess return is calculated as U it = R it R mt with R it is the return of firm i at period t while R mt is the market return at the same period. R it is measured by the price at the end minus the price at the beginning of the period under consideration. The difference is then divided by the price at the beginning of the period. R it = (P it P it )/P it The market return, on the other hand, is calculated by using data of the FT ALL Share Index. This calculation follows the model used by McDonald and Fisher (97). Such calculation is operated because it is difficult to calculate the excess return e it using CAPM in the equation.

6 R it = a I + b I R mt + e it The difficulty comes from the lack of data available for unseasoned issues, meaning that the calculation of intercept a I and coefficient b I is statistically not practicable. The excess returns are calculated as u N i= = N u it This study employs several types of period. Basically, all of these measures indicate short term excess returns. At first, this study employs four types of period to measure the excess returns between initial offering to a certain day. These four types of excess returns are measured between the initial offering to the first day of trading day, the first week of trading day, the first month of trading day, and the first year of trading day. These eccess returns are to measure the price performance enjoyed by investors who buy shares in initial offerings. Furthermore, this study also employ another three types of periods to measure the excess returns from the first trading day to the first week, the first month and this first year after. The purpose of employing these three types of periods concerns with the investors that do not have opportunity to buy stocks at offering but still hope to exploit immediate price movement in the after market. The uncertainty and underpricing It is hypothesised that underpricing represented by the excess return is related to the ex ante uncertainty. Among the five proxies for ex ante uncertainties, Beatty and Ritter (98) found two proxies as independent variables, a logarithm of one plus the number of uses of proceeds and reciprocal of gross proceeds. The use of the logarithm of one plus the uses of proceeds could causes a bias. To avoid, this study employ three proxies for the ex ante uncertainty. The first proxy is the leverages of the issuing firm. It is defined as the ratio of long term loans to total equity. It is believed that leverage could reflect the firm s risk. Also, this

7 could be the reason for a firm going public. Obviously, the proceeds can be used to restructure the capital by reducing the proportion of the loans to reduce the risk. The second proxy is the reciprocal of book value. This is a factor that investors use to perceive the ability of the firm to generate future earnings. And the last proxy is the reciprocal of the proceeds as applied by Beatty and Ritter. The relationship between the excess return and the price is U it = f (L, P, B) Where u it is the excess return of firm i at period t while L, P and B are the leverage ratio, the reciprocal of proceeds and the reciprocal of book value at time t respectively. It is expected to find positive relationships between the dependent and independent variables. On average, the higher the leverage, the larger is the risk of the company. A large portion of loans forces firms to provide a fixed amount of cash to pay the interest. If the economic condition is favourable, additional revenues will increase the earnings more steeply compared to firms without loans. But in an adverse situation, every decrease in revenue results in a sharper decrease in earnings. The book value and proceeds, on the other hand, reflect the ability of firms to sustain in the market and to invest more to generate earnings. It means that the larger the company, the smaller is the risk of failure. Thus, the reciprocals of both book value and proceeds should be related positively to the risks. That is why the regression coefficients of the three proxies are expected to be positive. DESCRIPTION OF DATA For the initial assessment, Table shows the classification of investment banking according to prestige. The grouping reflects the hierarchy representing the relative strength of the house in the very profitable area of underwriting. Table indicates that the income from underwriting contributes the second largest percentage and comes along way behind security commission. Table. Sources of gross income

8 Sources of incomes Percent of total Security commission.8% Underwriting 0.7% Trading 0.% Interest on customers debit balances 9.% Mutual Fund sale.9% Commodity commission.7% Others 8.% Note I) The data belong to New York Stock Exchange member firms. II) The table is adapted from Hayes Source: New York Stock Exchange research department Underwriting activity has special advantages for underwriters. First, underwriting commission serves as a defensive cushion during the bear markets. Second, underwriting could be a starting point for building up relationship with the issuers for other businesses. Further, the position of investment banking in the hierarchy has both shortterm and longterm implications for its profitability and competitive advantage to win more business of different types and size of service (Hayes, 97). In order to classify underwriters, researchers use three methods. The first was made by Hayes. This method uses tomstone and includes investment banks. He suggests some indicators that were/are used to evaluate their position in the hierarchy. Those are their past effectiveness in distributing issues; the size of firms, especially in negotiated issues; capital availability; reciprocal business; the repute of dominant partners; and number and quality of its corporate clients. The second method was created by Carter and Manaster (987). This method applies a continuous process to all tombstone ads in the investment Dealer s Digest from January 979 through December 98. The third was used by Johnson and Miller (988). This method employs a modification of the method created by Hayes. Those methods are adopted in this study, using the data of the UK market. The data for first assessment were collected from Extel Financial Ltd. This assessment faced three limitations in isolating the analysis for UK underwriters. First, London s International Securities Exchange is a place of international competition. Therefore, the prestige of a UK underwriter is not only determined by other UK underwriters but also by foreign peers. Second, it is difficult to find issuing firms using merely UK underwriters. And lastly, the syndicate of equity underwriting cannot be found in a certain periods. Instead, there were a lot of syndicates of liabilities underwriting.

9 Thus, the sampling is done as follows. First, underwriting of liabilities were chosen instead of underwriting of equity. This is definitely acceptable because the prestige is a general term applied for either liabilities or equity underwriting. Second, it is not necessary to choose UK underwriters as underwriting managers. The choice is based on whether UK underwriters are involved in those underwriting syndicates. As a result, the evaluation of prestige involves not only UK underwriters but also foreign underwriters. As a result, ninety nine UK underwriting firms were considered in this analysis. The data for the second and the third assessments, on the other hand, are focused on companies placing their stocks for the first time on the market. The main sources of information are data published by Extel Financial Ltd., Datastream, Microview, and Microextat. The related data are the oneyear stock price movements, book values, leverage ratios and the gross proceeds. The issuing companies that match the sampling method amounted to 97 companies employing 8 underwriting firms. Actually, there are other companies also placing their stock in the market. RESULTS The classification of underwriters The comparison of syndicates under the same managers indicates that only a few managers are likely to employ the same syndicates in term of membership and alphabetical order. Companies are categorised as special bracket firms if they are put under the manager or comanagers in the list before other firms. Concerning the underwriting led by Warburg (S.G) Sec, there is only one alphabetical list so it cannot be used to conclude whether they are special bracket, major bracket or sub major bracket firms. The pattern of extension and change in alphabetical order was commonly employed by most of the underwriting managers. Furthermore, Warburg (S.G) worked with three brackets of firms. The special bracket firms of Warburg s style were from Barclays de Zoete Wedd to the Royal Bank of Scotland plus IBJ International as an exception. The major bracket firms were Den Danske Bank, UBS Phillips & Drew Sec. And Westdeutsche Landesbank, while, the sub major bracket firms were from Banque Bruxelles Lambert S.A. to Sanwa International in the list. It is also a general pattern that there is no guarantee that sub underwriter will be treated in the same bracket by different managers. Warburg (S.G.) Sec as a sub underwriter seems to be treated as a sub

10 major bracket firm by a majority of managers. Nat West Capital Market Ltd., on the other hand, was treated as a special bracket firm by Warburg (S.G.) but as a major bracket firm by Hambros Bank. The treatment is influenced by the relationship and the ability of firms to negotiate with either managers or issuing firms. The relationship and ability to negotiate will result in the proportion of participation and agreement received from them. This is, as Hayes said, that to move a major bracket, a company needs confirmation from the important major and special bracket firms as social arbiters on Wall Street. Apart from those differences, there are two similarities among syndicates of different managers. First, managers tend to have one or two permanent partners in almost all underwriting under his management. For example, UBS Phillips & Drew Sec., as manager, almost always used Credit Suisse First Boston and frequently employs Warburg (S.G) as a sub underwriter. Second, UK underwriters tend to be placed in a similar position in many underwriting. Warburg (S.G), for example, was placed at bottom of the list by Nomura, Hambros Bank, Goldman Sachs, and UBS. To modify the Hayes model, the classification uses two steps. The first step is the use of the pattern of membership in certain brackets. If they have the opportunity to be a special bracket firm in an offering, they are classified as special bracket firms. This is because of the difficulty in getting the change to be a special bracket firm in a competitive market. The UK special bracket firms do not only compete with UK firms but also with foreign firms in either UK or European markets. This means that the inclusion of an underwriter in the special bracket is strongly based upon high prestige. Other underwriters that dominate major bracket positions are classified as major bracket firms. The rest of them are submajor bracket firms. The second step is the adjustment of the result above with the survey of the ranking of investment analysts by Extel Financial Ltd. Some firms may need to be removed to a certain bracket as the position from Extel is different from the Hayes model. For example, Kleinwort Benson is placed in the special bracket and Warburg is placed in the sub major bracket according to the Hayes model, Warburg is ranked in the first position above Kleinwort Benson. In this case, Warburb will be removed to the special bracket. The difference between the survey by Extel Financial and from the Hayes model exists because they use different criteria of ranking. The Hayes model emphasizes the extent to which underwriting firms occupy the business in the capital market, and the proportion of

11 participation in syndicates. This means that the Hayes model employs business results as the indicators. Extel Financial, on the other hand, employs the expertise of analysts from underwriting firms according to investment managers opinion as the indicator. The business result and the business expertise may not go hand in hand because the business results are determined not only by the expertise but also other factors such as relationship with clients, the marketing strategy, etc. Nevertheless, such adjustment is needed since the survey of Extel Financial represents the importance of underwriting firms. The use of this data is proposed to avoid the bias of issuers preference as mentioned in the preceding section and of the inclusion of a top ranked underwriter outside the special bracket. A special bracket firm may accept a sub underwriting regardless of the position or proportion of participation in a syndicate. Table : The distribution underwriters clients Code of Underwriter FTSE 8 OTHER ALPHA BETA OTHER TOTAL % OF TOTAL

12 9 0 Source: This is replicated from Quality of Markets, Companies Book 99, London Stock Exchange. Tables shows the small participation of sub major firms in offerings. Based on the size of issuing firms, it is found that under00 UK companies always employed prestigious underwriters as manager for their offerings. This is the reason that the changes of becoming managers for sub major bracket firms do not exist. Also, because the managers know the performance and ability of the underwriting of sub major firms, they are unlikely to be placed in special or major brackets. Further, as expected, the prestige of underwriters seems to apply both to liability and equity underwriting. Special bracket firms underwrite more numbers of either liabilities or equities than major brackets firms do and they underwrite more than sub major bracket firms do The level of underpricing Table provides the data of the means and standard deviations of the excess returns measured from the offerings. The data of mean indicate that, on average, investors could expect the excess returns of 0.7% from offerings. The excess returns, however, remain the same until the end of the first month from offerings and diminish by a half after one year. It could be expected that the excess returns would continue to decline afterward. Table : The means and standard deviations of the excess returns measured from the offering prices PERIOD SAMPLE MEAN STD Offerings to first day after market Offerings to first week after market Offerings to first month after market Total Prestigious underwriters Non prestigious underwriters Total Prestigious underwriters Non prestigious underwriters Total Prestigious 0.7% 0.0%.7% 0.0% 9.8%.% 0.% 9.77%.7%.90%.%.70%.7%.%.%.0%

13 Offerings to first year after market underwriters Non prestigious underwriters Total Prestigious underwriters Non prestigious underwriters.%.0%.%.9%.09% 0.%.99%.% Table also shows the difference between the excess return of issues underwritten by the prestigious and nonprestigious underwriters. In accordance with other studies, the underpricing levels of the former are less than the levels of the later. The reason is that the former tends to underwrite less risky companies. While the excess returns of offerings managed by the former are stable for only one month and those decrease very sharply after one year, the excess returns of offerings managed by the latter remain high for one year. Table gives the data of after market excess returns. The means of one week, one month and one year after the market are not significantly different from zero. This implies that the market adjusts the price very quickly within the period between offerings and the first day of trading. In other words, the price follows the principles of efficient market hypothesis afterward. This information tells the inventors not to buy stocks in the first trading day if they hope for excess returns. Table : The means and standard deviations of the excess returns of the total sample measured from the first day in the after markets PERIOD MEAN STD First day to first week after markets First day to first month after markets First day to first month after markets 0.% 0.% 0.0% 0.%.7%.70% The uncertainty and underpricing The following equations are the results of regression using the proxies, i.e. the reciprocal of both book values, the reciprocal of

14 proceeds, and the leverage ratio. By using a heteroskedasticity consistent covariance matrix to calculate standart errors and t statistics, the regression result with three independent variables is as follows Excess = * B * P 0.0 * L (.97) (.000) (.7) (.) with the figures in the parentheses are tstatistics While both the reciprocals of both book values and proceeds are positively correlated with the excess returns, as it is expected, the underpricing levels are negatively correlated with the leverages. This relationship is probably caused by the expectation of investors of gaining more earnings from loans (Stewart, 97). Earnings are positively correlated to the stock price. This is because the higher the earnings, the larger the expected dividends lifting the prices up. In good economic conditions, a firm may be able to produce a consistent pattern of growth of earnings per share by replacing equity with loans. In term of IPOs, this expected growth of earning through higher leverage results in a higher initial price thus lower underpricing. However, the tstatistics in the above equation demonstrates a very weak relationship between the excess return and the leverage. It is significant only at the 0 % level. The two other independent variables, on the other hand, are significant at the % level. So, the leverage as an independent variables is dropped. The result of the regression with two independent variables is as follows Excess = * B * P (.89) (.9) (.89) Note that R is very low, i.e. at 0.0. This is in accordance with the result of Beatty and Ritter s study. They justify that if R was high, it would imply that the actual initial return of an offering is predictable. The theory states that there is a positive relation between ex ante uncertainty and expected initial return. The reason for this positive relation is that it is difficult for investors to predict the actual initial return on high risk issue, giving rise to the winner s curse problem, even though the average initial

15 return in a large sample can be predicted with reasonable accuracy. The use of the regression, however, is very limited since the R is very low. Although on average the offerings are underpriced according to the regression, underwriters and investors cannot use the regression accurately to predict a certain offering. What investors can do is to diversify subscription to avoid any loss of overpricing. The low R may indicate that there are other better proxies able to explain the level of underpricing levels. I suspect they come from the fundamental factors of companies earnings. Further studies are certainly needed to find better predictive proxies. This study further evaluates the extent to which actual excess returns deviate from expected excess returns generated from the equation above. The deviations are indicated by the coefficients of constants and the independent variable, i.e. the expected excess returns. The three regressions show the following results. For total samples, the constant is not significantly different from zero while the coefficient of the expected excess returns is significantly equal to one. It means that the actual excess returns can predicted by and are the same as the expected excess returns. For IPOs managed by prestigious underwriters, the constant is significantly different from zero and the coefficient of the independent variable is.8. It implies that the actual excess returns deviate very much from the expected excess returns. For IPOs managed by nonprestigious underwriters, on the order hand, the constant is significantly not different from zero while the coefficient of independent variable is It means that the actual excess returns are less than the expected excess returns. The greater deviations of IPOs managed by prestigious underwriters may be caused by two possibilities. Firstly, because the greater expected excess returns indicate the greater risks, the underwriters underprice more for highlyrisky companies as the risk premiums. The underwriters are able to do so because of their reputation as the bargaining power. Also, prestigious underwriters are able to interfere the issues price to provide certain level of excess returns for investors. And the third possibility is that there is no or negative relationship between the prestige and the quality of information.

16 However, the prestigious underwriters still dominate the market in the period under considerations. It is not in accordance with the result of Beatty s study. It says that underwriters tend to lose the market share if IPOs managed by them have actual excess returns that deviate far from the excepted excess returns. The above result, on the other hand, indicates that nonprestigious underwriters performed better than prestigious underwriters in terms of the deviation of the quality of information. Yet, nonprestigious underwriters cannot outperformed the business of IPOs term of the market share or the number of offering managed by them. This probably indicates that there are other factors outside this study that can explain the underwriting performance. CONCLUSIONS Among the four categories of underwriting activities, underwriters notice the different importance of those activities for the business. The four categories are: determining the true value of issues, the pricing strategy, determining contracts, and advising timing and size of offerings. Underwriters are more concerned with determining the price of the issues and the true value of issuing firms. The reason is that both the jobs are strongly related to the performance. The accurate estimates of the true values enable the underwriter to underprice the issues correctly. The right price must be accepted by the issuers and investors. By doing so, the underwriters can maintain a good relationship with them for future business. The above study leads to the following five points. Firstly, the underwriters can be classified into three brackets using the modified Hayes model. The modification is applied because there is no syndicate employing only underwriters from one country. Also the modification is used to eliminate the bias of the Hayes model. Among 9 underwriters under analysis, there are 0 specialbracket firms, majorbracket firms and 8 sub major firms. Second, in accordance with other studies, this study also found underpricing in the UK market although it is still less than the underpricing of hot issue markets. The underpricing level, however, cannot be used to draw further conclusions about the market stage. Further analysis shows that different behaviour toward pricing exists among underwriters. Specifically, on average, prestigious underwriters underprice less than nonprestigious ones.

17 Third, the price movement follows the efficient market hypothesis immediately after the stocks being traded in the market. This is proved by the absence of excess returns in the after markets. Fourth, there are two proxies that are significantly related to ex ante uncertainty. Those are the reciprocals of both book value and proceeds. The proxies are positively related to the underpricing level. This is the same as the results of other studies proving the positive relationship between ex ante uncertainty and underpricing level. And lastly, by using the proxies to calculated the expected excess returns, it is found that actual excess returns deviate less from the expected excess returns for IPOs managed by nonprestigious underwriters than those managed by prestigious underwriters. The fact that nonprestigious underwriters cannot outperform prestigious underwriters in terms of the number of underwriting demonstrate that the deviations alone cannot explain the change in market share. There should be other factors explaining such change.

18 References. Beatty, Randolph P. and Jay R. Ritter, 98, Investment Banking, Reputations, and the underpricing of Initial Public Offerings, Journal of Financial Economics, pp... Carter, Richard B., 99, Underwriter Reputation and Repetitive Public Offerings, The Journal of Financial Research, Vol XV, No. (Winter 99), pp... Carter, Richard B. and S. Manaster, 987, Initial Public Offerings and Underwriter Reputation, Working Paper, University of Utah.. DeAngelo, H., and R. Masulis, 980, Optimal Capital Structure Under Corporate and Personal Taxation, Journal of Financial Economics 8, pp.9.. Hayes, III, Samuel L., 97, Investment Banking: Power Structure in Flux, Harvard Business Review, pp... James, C., 99, Relationship Specific Assets and the Pricing of Underwriter Services, Journal of Finance. 7. Johnson, James M. and Robert E. Miller, 988, Investment Banker Prestige and the Underpricing of Initial Public Offerings, Financial management (Summer 988), pp McDonald, J.G. and A.K. Fisher, 97, NewIssues Stock Price Behaviour, The Journal of Finance, pp Miller, Robert E. and Frank K., Reilly, 987, An Examination of Mispricing, Returns and uncertainty for initial public offerings, Financial Management, (Summer, 987), pp Neuberger, Brian M., and C.T. Hammond, 97, A Study of Underwriters Experience with Unseasoned New Issues, Journal of Finance and Quantitative Analysis, pp.77.. Ritter, Jay R., 989, Signalling and the Valuation of Unseasoned New Issues: A. Comment, The Journal of Finance, pp. 7.. Steward, Samuels S., 97, Corporate Financing, in Levine, Sumner N., Financial Analysts Handbook: Methods, Theory and Portfolio Management, pp Welch, Ivo, 989, Seasoned Offering, Imitation Costs, and the Underpricing of Initial Public Offerings, The Journal of Finance, Vol. XLIV, No. (June 989), pp.9.