臺灣博碩士論文加值系統

在國內，新上市股票一直最受股票投資人的青睞。但近兩年來，我們卻發
現到許多的新上 市股在上市之初便紛紛跌破承銷價，造成參加這些新上
市股公開承銷投資人的損失。因此 投資人該依循何種準則，在眾多的新
上市股中挑選出實務面上較不會跌破承銷價的新上市 股，來加以投資？
以及新上市股的"溢（折）價率"與各式的期間報酬率之間是否具顯著地
相關性？這些與投資人利害息息相關的課題，正是本文所欲加以探討的動
機所在。 本文的研究期間分成：模型建立期的民國76年至84年，與模型
測試期的民國85年1月至12 月兩段。首先，本文將先分別剖析兩段研究期
間內，新上市股跌破承銷價的行為，藉此 作為爾後建立跌破承銷價預測
模型的依據。其次我們將依國內股市結構、多空環境、產業 特性、與大
股東心態的不同，利用逐步迴歸分析法來建構出各式不同的新上市股跌破
承銷 價預測模型。並以民國85年為測試期，以檢驗各模型的測試效果。
最後，本文將挑選出預 測效果最佳的模型，在引入相關分析後，來進一
步探討新上市股的"溢（折）價率"與各式 期間報酬率間的相關性，以作
為投資大眾選擇新上市股進行投資時的參考依據.研究結果如下：1.民
國85年國內股市呈現出一個震盪走高的格局。2.民國85年新上市股跌破承
銷價的行為，多半發生在民國85年的空頭行情期間。而在3月中之後，新
上市股的"蜜月行情"，在多頭行情的助威之下，又有重新點燃之勢。3.本
文中所考慮的國內股市結構、多空環境、產業特性、與大股東心態等四種
模型建構屬性，在本文的測試期間（民國85年）內，確實能有效地掌控國
內新上市股跌破承銷價之行為。4.經由對民國85年新上市股所作的測試結
果，我們發現測試效果最佳的是模型一與模型七。因此，我們認為投資人
在掌控國內新上市股跌破承銷價行為上，可以將重點擺在下列兩項考量因
素上，即：Ⅰ.投資人若要以不同的切割期間來從事跌破承銷價預測模型
的建立，那麼拉長模型建立期的期間，將會對模型的預測效果產生正面且
顯著的幫助。Ⅱ.投資人若想更進一步地瞭解公司大股東的心態，則須勤
於瞭解各家新上市股公開說明書中，各項關於承銷價的財務報表資訊，藉
以判定新上市股的大股東對於公司股價所秉持的態度。因為經由本文可以
發現，大股東的心態，確實會影響新上市股上市後的股價表現。5.透過溢
（折）價率與各式期間報酬率之相關分析，我們認為投資人在投資國內新
上市股時，應該選擇投資那些屬於「溢價承銷」，並且溢價幅度高的新上
市股。

In Taiwan stock market,the newly listed stocks are the
investors'''' favorite targets owing to exist the"honey moon"
effect.However in the recent years,there were many newly listed
stocks broken down its initial public offering(IPO)prices .So
,the main objective of this paper is tring to find some
investing guideline through empirical study in order to handle
the behavior of investing in the newly listed stocks. The
reaserching period of this paper is divided into two periods.The
establishment period of forecasting model is from 1987 to
1995,and the testing preiod is the whole year of 1996.In this
study ,first we adopt "stepwise regression" to establish
forecasting models of the breakdown of IPO prices of the newly
listed stocks accroding to different attributions.Second, we
test the performance of seven forecasting models over the newly
listed stocks of 1996.Finally,we choice the best model to
analyze the relations between the rate of premium(discount) of
IPO prices and the different kinds of duration return.The main
resurts of this paper include:1.The newly listed stocks IPO
prices become more easy to breakdown duringthe bearish period of
1996.After the middle of march , the"honey moon"effect restarted
under the stimulus of bullish market.2. We have contemplate four
different attributions in the establishment of forecasting
models.Those attributions can effectively capture the behavior
of the breakdown of IPO prices in our testing period.3.Accroding
to the testing results of seven forecasting models,the best
modles are the modle 1 and modle 7.So, when investors want to
invest in the newly listed stocks,we recommend that the key
points should be the following two factors：a.If investors try
to establish forecasting models of the breakdown of IPO prices
with different duration,to extend the duration will increse the
predictive power of the econometric model.b.Investors must to
perceive the finacial statement about the formula in
establishing IPO prices in public issue communigue .Investors
can pierce the powerful stockholders attitude to IPO price which
will affect the supportability of price of the newly listed
stocks.4. Accroding to the result of correlation coefficient
analysis,we recommendinvestors should invest in those newly
listed stocks with higher degree of premium of IPO prices .