台灣痛苦指數的檢驗與修正

Empirical Validation and Revision of the Misery Index for Taiwan

陳韻竹(Yun-Chu Chen)

指導教授 ： 謝德宗

國立臺灣大學/社會科學院/國家發展研究所/碩士(2010年)

https://doi.org/10.6342/NTU.2010.10018

摘要

本研究旨在探討民生痛苦指數與金融痛苦指數的計算公式在實證上是否符合台灣國情，最後並提出台灣適用的痛苦指數修正公式。
依據國內外文獻對痛苦指數的評論與修正，首先以1990年第一季~ 2009年第二季的台灣資料分別檢驗民生痛苦指數、金融痛苦指數與國際重大事件是否有時間上的關聯性，其次再檢驗1999年第一季 ~ 2009年第二季的台灣失業率、通膨率，與股價指數變動率、間接匯率（的）變動率、經濟成長率之各種變數組合對台灣消費者信心指數的解釋力，以找出具有最佳解釋力的組合，並依此提出痛苦指數在台灣的修正式。
實證結果與分析如下：
一、綜合各結構性轉檢定結果來看，民生與金融痛苦指數對國際重大事件的發生幾無預測或解釋能力，因此痛苦指數的計算公式需要被修正以符合台灣民情。
二、提出台灣適用的痛苦指數修正式：
將台灣消費者信心指數作為人民福祉的代理變數，並以多元迴歸分析方法個別探討各種總體變數的組合對台灣消費者信心指數的解釋力，藉此找出具有最佳解釋力的變數組合，並獲致以下結果：
（一）將失業率與通膨率拆成同一模型的兩自變數後再對消費者信心指數進行迴歸的估計結果比失業率與通膨率直接相加（即民生痛苦指數）後對消費者信心指數進行迴歸的估計結果更具有解釋能力；金融痛苦指數的實證結果亦然。亦即民生痛苦指數與金融痛苦指數的複合變數定義在台灣的實證檢驗結果不佳，需要修正。
（二）當失業率或通膨率降低，或股價指數變動率、間接匯率變動率、或經濟成長率上揚時，皆使台灣消費者信心指數提升，符合理論預期。
（三）綜合比較各模型迴歸估計結果可得：
1. 在台灣各總體變數中，失業率、通膨率與股價指數變動率對台灣消費者信心指數有顯著的影響。
2. 由所達之顯著水準可知，各變數對消費者信心指數的影響之顯著性大小依序為股價指數變動率與經濟成長率、通膨率、失業率，最後是間接匯率變動率。
3. 由迴歸係數估計值可知，失業率為所有變數當中對消費者信心指數影響幅度最大的變數，其次為通膨率，再其次為股價指數變動率。
4. 從迴歸最佳估計式的結果可得，失業率、通膨率與股價指數變動率的係數比大約為26：9：（-1）。亦即，失業率對消費者信心指數的影響幅度約為股價指數變動率對台灣消費者信心指數影響幅度的近26倍，通膨率對台灣消費者信心指數的影響幅度是股價指數變動率對台灣消費者信心指數的影響幅度的近9倍，而失業率對台灣消費者信心指數的影響幅度是通膨率影響幅度的近2.7倍。由此可知，民生痛苦指數的計算公式裡的失業率係數權重確實是被低估。
（四）依前述實證結果提出台灣適用的痛苦指數修正公式。最後，將此「修正版」痛苦指數再進行迴歸估計以及結構性轉變檢定兩種檢驗，並獲致結論如下：「修正版」痛苦指數對於台灣人民的經濟痛苦程度確實具有良好解釋能力。換言之，當「修正版」痛苦指數上升，將使台灣消費者信心指數降低，人民經濟痛苦程度愈大。

關鍵字

痛苦指數 ； 民生痛苦指數 ； 金融痛苦指數 ； 結構性轉變 ； 多元迴歸 ； 台灣消費者信心指數 ； 福祉 ； 生活滿意度

並列摘要

This paper empirically investigated the applicability of the misery index and the financial definition of the misery index to Taiwan and finally proposed a revised equation of the misery index for Taiwan.
Based on the views and modifications of the misery index in previous literature, this paper used the data in Taiwan from 1990Q1 to 2009Q2 to first examine the correlation between the misery index and worldwide significant events as well as the correlation between the financial definition of the misery index and worldwide significant events, respectively. Later, this paper examined the explanatory power of various combinations of variables, including unemployment rate, inflation rate, the rate of change of stock indices, the rate of change of the indirect exchange rate, and economic growth rate for Taiwan’s Consumer Confidence Index (Taiwan’s CCI) from 1990Q1 to 2009Q2. By finding the combination of variables that best explains Taiwan’s CCI, this study attempted to revise the equation of the misery index for Taiwan.
Empirical results and analyses are as follows:
I. Through various tests of structural changes, this paper found that the misery index and the financial definition of the misery index cannot predict or explain the effects of worldwide significant events. Therefore, the equation of the misery index should be rectified according to practical conditions in Taiwan.
II. Formulation of a modified equation of the misery index for Taiwan:
This paper used Taiwan’s CCI as a proxy variable of well-being and applied multiple regression method to investigate the explanatory power of different combinations of variables associated with the misery index for Taiwan’s CCI to find the combination that best explains Taiwan’s CCI. The results are as follows:
1. The model with unemployment rate and inflation rate divided as two independent variables can better explain Taiwan’s CCI than the model with unemployment rate and inflation rate directly summed as one composite variable (i.e. misery index). The same conclusion was also observed in the test of the financial definition of the misery index. In other words, the composite variables used in the misery index and the financial definition of the misery index cannot adequately capture the misery in Taiwan in empirical validation and should be modified.

2. Decline in unemployment rate or inflation rate or any rise in the rate of change of stock indices, the rate of change of the indirect exchange rate, and economic growth can all lead to a rise of Taiwan’s CCI as theoretically expected.

3. Findings from model regression estimations include:
(1) Among all macroeconomic variables, unemployment rate, inflation rate, and the rate of change of stock indices have significant effects on Taiwan’s CCI.
(2) In terms of significance level, Taiwan’s CCI is most significantly influenced by the rate of change of stock indices and economic growth rate, followed by inflation rate, unemployment rate, the effects of the rate of change of the indirect exchange rate are least significant.
(3) In terms of the magnitude of the regression coefficients, unemployment rate, among all the variables considered, has the largest magnitude of influence on Taiwan’s CCI. Inflation rate has the second largest magnitude, followed by the rate of change of stock indices.
(4) According to the best estimator in regression, unemployment rate, inflation rate, and the rate of change of stock indices are in a ratio of approximately 26:9:(-1). In other words, unemployment rate is nearly 26 times more influential to Taiwan’s CCI than the rate of change of stock indices to Taiwan’s CCI; inflation rate is nearly 9 times more influential to Taiwan’s CCI than the rate of change of stock indices to Taiwan’s CCI; unemployment rate is nearly 2.7 times more influential to Taiwan’s CCI than inflation rate to Taiwan’s CCI; It can be inferred that the weight of unemployment in the misery index had been underestimated.

4. Based on the above empirical results, this study proposed a revised equation of the misery index for Taiwan. Finally, the “revised” misery index was validated through regression estimations and tests for structural changes to derive the following conclusion: The “revised” misery index indeed has a good explanatory power for the economic misery in Taiwan. In other words, the rise of the “revised” misery index will induce a significant decline in Taiwan’s CCI and a higher level of economic misery.

並列關鍵字

misery index ； financial definition of the misery indexs ； structural change ； multiple regression ； Consumer Confidence Index (CCI) ； well-being ； life satisfaction

參考文獻

張職揚（2005），「檢驗結構性轉變在網路產品之影響—Chow, CUSUM, STAR方法之比較」，國立中山大學經濟學研究所。

吳津苗（2007），「台灣、美國總經月數據與台股股價指數之關聯性」，中央大學產業經濟研究所碩士論文。

張偉諒（2007），「消費者信心指數與台灣股票市場相關性之研究」，中央大學財務金融系碩士在職專班碩士論文。

Andrews, D. (1993), “Tests for Parameter Instability and Structural Change with Unknown Change Point,” Econometrica, 61(4), pp.821-856. Retrieved December 30, 2009, from EconLit with Full Text database.

Barro, Robert J. (1996), Getting it right : markets and choices in a free society, Cambridge, Mass. ; London, England.