本研究旨在探討競價拍賣承銷新制是否能有效改善新股上市（櫃）初期鉅額價差之異狀，所使用的理論架構與方法主要為『訊息傳遞理論』。為充分瞭解研究中所提出之問題，本論文首先彙集整理國內外與新股上市價差有關之文獻，據以為研究假設形成之理論基礎，次就我國承銷制度的變革經過與目前競價拍賣制在台灣的實際運作情況作重點介紹與說明，俾利瞭解我國現行承銷制度之全貌。實證部份先對競價公司樣本進行屬性分析，再比較不同制度下兩組樣本上市（櫃）後股價超常報酬是否存在顯著差異，並提出影響實證結果的可能因素，以判斷現行競價拍賣制度是否足以改善蜜月期之異常股價行為。

依據訊息傳遞理論與現行競價拍賣之制度特性，本研究臆測若發行公司之現況良好且對未來營運充滿信心，則會有動機去發展出一些訊息機制(mechanisms)傳遞予投資大眾，而承銷價低估即為訊息之一。在非績優公司無意傳遞此等訊息、而競價拍賣又較易形成相對偏高承銷價之假設下，推論發行公司董監事持股比例高、總資產規模較小（風險程度較高）及股價表現不具特殊題裁之新上市（櫃）公司較傾向於透過競價拍賣制公開承銷新股。再就資訊不對稱經濟學論之，當前我國證劵發行市場之癥結除了承銷制度設計不當外，尚包括市場競爭性不足、財務資訊環境不佳等，新版競價拍賣制固然能使承銷價之決定方式透過市場機能反映供需，理論上承銷價當已接近市價，然在整體財務資訊環境與股票市場未臻合理完備之際，本研究認為僅憑制度面的變革並不足以解決承銷體系長期存在的弊端。

新版承銷制之運作概況及上述假說經實證檢定後獲致以下結論：

一、競價拍賣制之實施概況

(一) 截至 86 年 12 月 31 日止共有 31 家發行公司透過競價拍賣方式公開承銷新股，競價承銷總金額累計達新台幣 21,392,037,350 元。而全體競價公司樣本之平均得標率為 23.22%，多數個案的公開承銷價達投標底價 1.5 倍之最高水準，顯示投標競爭狀況十分激烈。

(二) 法人參與投標比例約為 25%，表示目前競價拍賣之投標市場尚未由機構法人所主導；而法人平均得標率為 18.67%，顯示法人所設定之新股合理價位仍較一般投資人為保守。

(三) 在進行相關係數之統計檢定後發現，競價樣本投標部份的熱絡程度會對投資人參與抽籤之意願形成正面影響。

二、競價公司樣本屬性推論

『上櫃公司中規模較小、風險程度較高者』及『所屬產業型態為傳統產業』是目前競價公司的重要屬性，『董監事持股比例高』與『掛牌於店頭市場』則未獲得實際數據的支持。

三、不同制度下的股價績效

(一)我國新上市（櫃）股票自上市（櫃）後第 1 個交易日至第 300 個交易日之間存在顯著正向超常報酬，表示整體而言，新股承銷價格有持續低估之現象。

(二)競價拍賣制縮短了新股蜜月期持續天數，然在新舊制度下，兩組樣本上市 （櫃）後第 1 個交易日至第 22 個交易日之超常報酬並無顯著差異；至第 23 個交易日起，競價公司股價之超常報酬方顯著的小於非競價公司樣本，顯示競價拍賣承銷新制改善新股蜜月期價差之成效未盡理想，且造成中長期股價行為較劇烈的波動。

承上，本研究認為就承銷體系長期存在的弊端而言，僅憑承銷制度的變革並不足以達到治本之功效，此一結論可供主管機關作為制定相關決策時之參考。

The objective of this research is to study whether the new system, auction procedure, can effectively improve the IPOs underpricing during the honeymoon period. The theory applied in this research is mainly “Signaling model”.

For the purposes of realizing the problems arising from this study, first, the relative reseach documentation and references were collected and investigated. Second, the existing and new systems of underwriting applied in Taiwan were introduced and anaylyzed. According to these two approaches, we can have fully understanding on both the old and the new systems. On the aspect of the empirical test, first, the characteristic analyses of auction samples were conducted. Second, the comparison of whether there were significant different of abnormal returns of common stocks between two sets of systems was made. Then the possible factors which affect the results of the empirical test were found, thus could we judge whether the new system can improved the fact of abnormal price performance of common stocks new issued during the honeymoon period.

Based on the signaling theory and the characteristics of the new system, the following assumptions were made:

1.If the IPO firms operate well are confident of the future performance, they will be willing to transfer signals (e.g. underpricing) to their potential investors. However, the other IPO firms don't have that incentive to do so. Therefore, we expected the firms which their inside equity ratios are higher, the total assets are less, etc, will prefer to issue common stocks through the auction procedure, because that the new system inclines to form relative higher issue prices.

2.According to the Asymmetric-Information Hypothesis, in addition to the imperfect underwriting system, the existing problems of the IPO market are insufficient competency, undeveloped financial information environment, etc. Even the new system seems to reflect the demand and supply circumstances of the market, we will believed that the change of underwriting system is not sufficient to improve the existing defects of the IPO market.

The following conclusions were reached according to this study:

1.The operation of the new system:

(1)There were 31 firms applied the new system during the period of this study. The total underwriting amounts are NT$21,392,037,350.

(2)The involvement of the insitutional investors occupied 25% of the total offerings.

(3)According to the relative coefficient analysis, the situation of auction positively affected the situation of drawing.

2.The comparison of stock price performance between the different underwriting systems:

(1)There are positive abnormal returns on the new issued common stocks which go on for 300 or more dealing days.

(2)The new system shortens the length of the honeymoon periods. However, The abnormal returns had no significant difference between two samples of the old and new systems from the 1^st to the 22^nd dealing day. The abnormal returns of new system samples would be significantly less than that of the old system samples from the 23^rd dealing day. This shows that the new system doesn't perform its objective which is to correct the underpricing of the old system well.

According to the conclusions made above, we believed that the change of underwriting system is not sufficient to improve the existing defects of the IPO market. This result of the research can be taken into consideration when making relative decisions.