臺灣博碩士論文加值系統

本文建構一個縱橫平滑轉換迴歸模型 (panel smooth transition regression model
PSTR)，以探討在不同的景氣對策訊號分數下，各未預期財務指標 (unexpected financial
indices) 對股價報酬的非線性衝擊效果。所選用的未預期財務指標，包括：未預期負債
比率、未預期稅後淨值報酬率與未預期稅後淨利率，分別代表償債能力、經營能力與獲
利能力的曝險 (exposure) 程度。實證上，針對台灣 29 家上市 (櫃) 金融公司 (含金控公
司與一般商業銀行) 在 2008 年第 1 季至 2017 年第 2 季的縱橫資料(panel data) 進
行估計。共計有 1102 筆觀察值。經由 PSTR 模型的估計結果，可提供更正確的股價報酬
值，以作為公司進行購併、投資者篩選投資標的，以及政府穩定股市的參考依據。
本研究的實證結果，歸納如下：
1. 未預期財務指標 (未預期負債比率、未預期稅後淨值報酬率與未預期稅後淨利率)
變動對股價報酬的影響，決定於各期景氣指標值。兩個景氣對策訊號分數之門檻值
(22.4748 與 31.0013)，將非預期財務指標對股價報酬的影響區分為三個區間。換言之，
股價的曝險程度，是隨時間而變動的 (time-varying)，亦即各期的景氣指標分數扮演著關
鍵角色。
2. 當景氣對策訊號分數在較高水準 (大於門檻值 31.0013) 或較低水準 (低於門檻
值 22.4748) 時，未預期負債比率提升，不利於股價報酬上揚，且由兩門檻形成區間之正
貢獻值轉為負貢獻值。其理由在於投資者擔心景氣指標低於 22.4748 時而持續惡化，或
景氣指標大於 31.0013 時而可能反轉，故未預期負債比率上升，不利於股價；惟其影響
不大，可能因為金融業本身屬於高負債比行業，故在本實證結果中為不顯著。
3. 當景氣對策訊號分數為門檻值 ( 22.4748 與 31.0013) 時，未預期稅後淨利率對
於股價報酬的貢獻為 -0.0015。當景氣對策訊號分數介於門檻值之間時，未預期稅後淨
利率的貢獻由 -0.0030 逐漸降為 -0.0015。當景氣對策訊號分數大於 31.0013 或小於
22.4748 時，未預期稅後淨利率對股價報酬的衝擊接近於零。當景氣位於穩定的「綠燈」
( 23-32 分) 時，投資人擔心未來景氣是否進入黃紅燈或黃藍燈的階段，故非預期稅後淨


II
利率提升，不利於股價報酬。然而當景氣經較佳或較差的階段移動時，因為景氣逐漸明
顯轉往好或壞的方向移動，故該負向衝擊逐步縮減。
4. 當景氣對策訊號分數為 26.7381 時，未預期稅後淨值報酬率對股價報酬的影響為
0.0536。當景氣對策訊號分數為 22.4748 或 31.0013 時，該影響降為 0.02385。此外，當
景氣對策訊號分數大於 31.0013 或小於 22.4748 時，該影響轉為負值 (-0.0059)。實務上，
當景氣對策訊號分數介於 23-31 分時，代表景氣穩定的「綠燈」，故未預期稅後淨值報酬
率增加，可提升股價報酬。然而，隨著景氣移往不佳的黃藍燈 (17-22)，或移往活絡的黃
紅燈 (32-37 分) 時，該正向的貢獻逐漸下降，甚且較為負向的。其理由與未預期負債比
率對股價報酬的影響類似。
本研究的政策建議如下：
1. 投資者、金融業者與政府金管單位在評估各財務指標的曝險程度及進行相關的避
險措施時，應隨著各期的景氣變化，作逐期的調整，而非如傳統線性模型所估計的固定
曝險係數，並錯估股價報酬與避險措施。
2. 就投資權益與增資舉債的時機而言，在景氣燈號為綠燈 (23-32 分) 時，意外的
舉債 (負債比提升) 與意外的股東權益上升 (稅後淨值報酬率)，有利於股價報酬的提昇。
反之，當景氣燈號分數大於 32 或小於 23 時，非預期買回庫藏股，有助於提升股價報
酬。
3. 當景氣對策訊號分數在較高水準 (大於門檻值 31.0013) 或較低水準 (低於門檻
值 22.4748) 時，三種財務指標曝險程度增加，均不利於股價報酬。政府金管當局宜監控
並採取因應措施，以防外在總體經濟環境變化帶來景氣可能持續惡化 (低景氣分數) 或
反轉 (高景氣分數) 對金融股與金融市場所帶來的衝擊。

This paper constructs a panel smooth transition regression model (PSTR) to explore the
nonlinear impact of unexpected the financials indices on stock returns under different scores of
in monitoring indicator. The unanticipated financial indicators selected include unexpected debt
ratio, unanticipated return on equity, and unexpected net profit margin, which represent the
degree of exposure of debt-paying ability, operating capacity, and profitability, respectively.
Empirically, we estimate the panel data of Taiwan''s 29 listed financial companies (including
financial holding companies and commercial banks) from 2008:1Q to 2017:2Q, a total of 1102
observation. The PSTR model, can provide a more accurate estimation results of stock returns
for the companies to conduct mergers and acquisitions, investors to select investment targets,
and the government to stabilize the stock market.
The empirical results are summarized as follows:
First, the impacts of unanticipated financial indicators (uninformed debt ratio, unanticipated net
return on equity, and unanticipated net profit margin) on share return are determined by each
period’s monitoring indicators. Two estimated thresholds of the monitoring indicator (22.4748
and 31.0013) divide the impact of unintended financial indicators on stock returns into three
regimes. In other words, the degree of exposure to stock prices is time-varying, implying that
the monitoring indicators play a key role in the stock returns and risk premiums.
Second, when the scores of the monitoring indicator are at a higher level (above the threshold


IV
of 31.0013) or a lower level (below the threshold of 22.4748), the increase in the unanticipated
debt ratio would be harmful to the stock returns. The reason is that investors worry about the
continuous deterioration the monitoring indicator below 22.4748, or the possible reversal of the
monitoring indicator above 31.0013. Therefore, the increase in the unexpected debt ratio will
pall dow stock returns. However, the impact is not much due to the characteristic of high debt
ratio in the financial companies.
Third, the contribution of the unexpected net profit margin to stock returns is -0.0015.When the
monitoring indicator reaches the thresholds (22.4748 and 31.0013).The contribution runs from
-0.0030 to -0.0015 when the monitoring indicator is in the interval [22.4748,31.0013].The
contribution contribution approaches to gro as the monitoring indicator is above 31.0013 or
below 22.4748.When the economy is in a steady “green light” (23-32 points), investors are
worried about whether the prospect of the economy will enter into the stage of a yellow light
or yellow blue light, so the increase in the net profit margin is harmful to stock returns. However,
when the economy moves through a better or worse stage, the negative impact gradually
weakens because the business gradually shifts to a good or bad direction.
Fourth, when the monitoring indicator 26.7381, the influence of the unexpected on stock returns
is 0.0536. When the indicator is 22.4748 or 31.0013, the impact falls down to 0.02385. In
addition, when the score of the indicator is greater than 31.0013 or less than 22.4748, the effect
turn out to be negative (-0.0059). In practice, when the indicator locates in the interval
(23.31),the economy is in a stable stape. Therefore, the indicator in the unexpected return on


V
equity can pash up stock returns. However, as the economy moves toward the poor yellow-blue
lights (17-22,points) or toward the active yellow-red lights (32-37 points), the positive
contribution gradually disappears and even becomes negative. The reason is similar to the
impact of the unanticipated debt ratio on stock returns.
The associated policy suggestions include:
First, investors, financial managers and government regulators should assess the exposure of
various financial indicators and perform relevant hedging strategies period by period according
the change in the monitoring indicator. Conversely, the constant exposure coefficients obtained
from the traditional linear model will result in biased hedging strategies.
Second, in terms of the timing of equity-issuance and debt-financing, the increases in the
unexpected debt ratio and unexpected return on equity are favorable to stock returns. As the
monitoring indicator displays the “green light ”(23-32 points). Contrary, the effect is negative
as the indicator is the above 32 or below 23.
Third, when the monitoring indicator at a higher level (above the threshold 31.0013) or at a
lower level (below the threshold 22.4748), the increase in the exposure of the three financial
indicators increases, is unfavorable for stock price returns. It is advisable for the financial
authority to monitor and adopt corresponding strategies to prevent the negative impact of the
external macroeconomic environment on stock returns and stock market through continuously
deteriorating the monitoring indicator or reversing the boom.

摘要.............................................................................................................................................I
Abstract.....................................................................................................................................III
致謝........................................................................................................................................ VII
表目錄......................................................................................................................................IX
圖目錄.......................................................................................................................................X
第壹章 緒論..............................................................................................................................1
第一節 研究背景與動機..................................................................................................1
第二節 研究目的..............................................................................................................4
第三節 研究架構與流程..................................................................................................5
第貳章 文獻回顧......................................................................................................................7
第一節 股價報酬率相關文獻..........................................................................................7
第二節 獲利風險或未預期的獲利之相關文獻............................................................10
第三節 縱橫平滑轉換迴歸模型之相關文獻................................................................12
第參章 實證模型....................................................................................................................14
第一節 線性自我迴歸模型............................................................................................14
第二節 縱橫平滑轉換迴歸模型....................................................................................15
第肆章 研究方法....................................................................................................................19
第一節 縱橫單根檢定....................................................................................................19
第二節 線性檢定與無剩餘非線性檢定........................................................................24
第二節 無剩餘非線性檢定............................................................................................26
第伍章 實證結果與分析........................................................................................................28
第一節 資料來源與處理................................................................................................28
第二節 敘述統計分析....................................................................................................30
第三節 縱橫單根檢定結果............................................................................................32
第四節 線性動態迴歸模型............................................................................................33
第五節 線性檢定............................................................................................................35
第六節 無剩餘非線性檢定............................................................................................36
第七節 PSTAR 估計結果分析.......................................................................................38
第陸章 結論與建議................................................................................................................42
第一節 結論....................................................................................................................42
第二節 政策建議............................................................................................................44
參考文獻..................................................................................................................................45



表目錄
表 5-1 樣本金融公司.........................................................................................................28
表 5-2 資料來源及衡量方法.............................................................................................29
表 5-3 敘述統計分析.........................................................................................................31
表 5-4 縱橫單根檢定結果.................................................................................................32
表 5-5 線性動態迴歸模型估計結果.................................................................................34
表 5-6 線性檢定結果.........................................................................................................35
表 5-7 無剩餘非線性檢定結果.........................................................................................36
表 5-8 PSTR 模型估計結果.............................................................................................41




圖目錄
圖 1-1 研究流程架構 ..........................................................................................................6
圖 3-1 羅吉斯型轉換函數 (m=1) ....................................................................................17
圖 3-2 指數型轉換函數 (m=2) ........................................................................................18

李嵩萍，(2004)。經營效率對公司價值與股價之影響－以 TFT-LCD 產業為例。國立高
雄第一科技大學金融營運所碩士論文。
柯淑珍，(2010)。財報資訊與股票報酬率之分量迴歸分析。國立高雄應用科技大學金融
資訊研究所碩士論文。
呂麒麟、洪嘉聲、周至文，(2008)。股權結構對資訊揭露在每股盈餘變動下之影響：以
台灣上市公司為例。致遠管理學院學報，第 3 卷第 2 期：53-80。
謝佩芬，(2004)。財務指標與非會計資訊對股價之影響－Ohlson 模型之延伸應用。中
原大學國際貿易研究所學位論文。
楊姝涵，(2011)。總體經濟指標及財務指標對股價報酬的影響-台灣銀行業與證券業之
比較研究。臺北大學國際財務金融碩士在職專班學位論文。
劉肇元，(2011)。機構投資人持股比率對公司績效與股價報酬影響之研究－以上市電子
股為例。國立高雄第一科技大學金融系碩士論文。
辜禹涵，(2015)。信用交易與匯率波動對股價報酬率的影響。國立高雄應用科技大學國
際企業研究所碩士論文。
陳建勝、陳美菁、林明宏，(2004)。我國產險業資本結構與風險對獲利能力影響之研
究。風險管理學報，第 6 卷第 3 期：273-290。
呂青樺、沈中華，(2006)。影響大陸地區銀行獲利能力與風險的因素。中國大陸研究，
第 49 卷第 3 期：119-142。
劉景中，(2008)。銀行集中度與台灣銀行業的獲利性及風險。經濟論文叢刊，第 36 卷
第 3 期：327-355。
陳富強、馮淑琴、方顯光，(2011)。探討台灣地區本國銀行獲利能力與市場能力－二階
段 DEA。華人經濟研究，第 9 卷第 2 期：33-52。
Avkiran., Rowands. (2008). How to better identify the true managerial performance: State of
the art using DEA. Special Issue on Knowledge Management and Organizational
Learning. 36.317-324.
Bens, D.A., V. Nagar., D.J.Skinner., M-H. F.Wong. (2003). Employee stock options, EPS
dilution, and stock repurchases. Journal of accounting & economics,36,51-90.
Bessec, M., Fouquau,J. (2008). The non-linear link between electricity consumption and
temperature in Europe: A threshold panel approach. Energy Economics, 30(5): 2705
2721.
Colletaz, G., Hurlin, C. (2006). Threshold wffects of the public capital productivity: An
international panel smooth transition approach. University of Orleans Working Paper.
Cheng,C. H., Wu, P. C. (2013). Nonlinear earnings persistence. International Review of
Economics & Finance, 25, 156-168.
Fouquau, J., Hurlin, C.,Rabaud, I. (2008). The feldstein–horioka puzzle: A panel smooth
transition regression approach. Economic Modeling, 25, 284-299.
Fried et al., (2002). Graphical models for multivariate time series from intensive care
monitoring. Statistics in Medicine, 21, 2685-2701.
Granger C. W. J., Teräsvirta, T. (1993). Modelling non-linear economic relationships.
Research Institute of the Finnish Economy.
González, A., Teräsvirta, T., van Dijk, D. (2004). Panel smooth transition regression model
and an application to investment under credit constraint, Working Paper. Stockholm
School of Economics.
González, A., Teräsvirta, T., van Dijk, D. (2005). Panel smooth transition regression models.
Research paper 165. University of Technology, Sidney: Quantitative Finance Research
Centre.
Hansen, B. E. (1999). Threshold effects in non-dynamic panels: Estimation, Testing and
inference. Journal of Econometrics, 93, 345-368.
La Porta, R., Lopez-De-Silanes, F., Shleifer, A. (2002). Government ownership of banks. The
journal of finance, 57(1), 265-301.
Monaheng, Seleteng., Manoel, Bittencourt., Reneévan, Eyden. (2012) .Non-linearities in
inflation–growth nexus in the SADC region: A panel smooth transition regression
approach. Economic Modelling, 30, 149-156.
Pan, S. C., Liu, S. Y., Wu, P. C. (2014). Re-testing the tourism-led growth hypothesis using
panel smooth transition regression models. Tourism Economics, 20, 39-50.
Teräsvirta, T. (1994). Specification, estimation and evaluation of smooth transition
autoregressive models. Journal of American Statistical Association, 89: 208-218.
Wu, P. C., Liu, S. Y., Pan, S. C. (2013). Nonlinear bilateral trade balance-fundamentals nexus:
A panel smooth transition regression approach. International Review of Economics and
Finance, 27, 318-329.