台灣地區自僱家戶低報所得推估與真實所得分配之探討

The Estimation of Self-Employment Income Underreporting and the Evaluation of Real Income Distribution in Taiwan

謝婷伃(Ting-Yu Hsieh)

指導教授 ： 林恭正

共同指導教授 ： 蘇建榮(Jain-Rong Su)

國立臺北大學/公共事務學院/財政學系/碩士(2013年)

摘要

家庭收支調查資料中的所得資料不僅為政府制定相關社會福利及公開資訊數字的依據，也經常使用於學術研究，因此家庭收支調查資料能否真實反映家戶所得情況非常重要，曾巨威(2010)曾提及的家庭收支調查資料樣本數不足，代表性也有待商榷，如果樣本中所得資料的正確性也有問題，則據以推估的所得分配指標將無法真實反映台灣所得分配的情況。然過去台灣對這方面研究不足，僅有呂逸晨(2011)以支出基礎法探討2005 年及2009 年家庭收支調查資料中低報所得的情況，並未進一步利用其調整後所得計算真實的所得分配，用以分析台灣地區真實所得的分配情況。
本研究使用支出基礎法對2001年至2010年的資料進行估算低報所得的程度，並使用調整後的真實所得推估所得分配指標，分別重估五等分位差距倍數及吉尼係數指標，並與調整前的指標比較。
結果顯示10 年間平均低報所得4%到36%，支持我們對於自僱者家戶相對受僱者家戶而言會隱瞞其真實收入的假設，10 年來所得低報總數高達上百億元，每年平均每戶低報25,500 元。研究結果顯示所得最高者及最低者較容易低報所得，白領工作人員比藍領工作人員容易低報所得。調整後所得分配更為不均，五等分位差距倍數及吉尼係數分別由4.91 擴大為5.00 及0.307 增加至0.313，可以肯定所得低報的存在與真實所得分配較申報所得更為不均。

關鍵字

所得短報、台灣地區家庭收支調查資料、支出基礎法、五等分位差距倍 數、吉尼係數

並列摘要

The data of The Survey of Family Income and Expenditure (SFIE) of Taiwan is heavily used to establishing government welfare policies as well as studying the income distribution of Taiwan in many academic researches. However, Zeng(2010) posed two questions about the SFIE data: one is the insufficient samples of SFIE and the other is the representation of samples. Therefore, it is very important to recognize whether the SFIE data reflects the real Taiwan’s household income or not.
In a previous study by Lu (2011), he estimated the degree of underreporting income of SFIE based on the expenditure-based approach for the years of 2005 and 2009. However, Lu did not use the revised income data to analyze the real income distribution in Taiwan.
This study extends the work of Lu (2011) to examine the underreporting income from 2001 to 2010 and further use the revised real income data to calculate the index of income distribution in Taiwan. Both the index of Gini coefficient and the ratio of top 20% to lowest 20% income are estimated and compared with the index of unrevised income distribution.
The result shows that the average rate of underreporting income is from 4% to 36% during the 10 years period which is consistent with our hypothesis that compared to the employed households, self-employed households are more likely to conceal their real income. The estimated underreporting incomes were more than ten billion NT. dollars within 10 years, which means every household averagely underreported NT. 25,500. The results also show that the rich and the white-collar are more likely to underreport their income than the poor and the blue-collar do, respectively. Finally, the distribution of adjusted income is more unequal than that of reported income. The ratio of top 20% to lowest 20% is increased from 4.91 to 5.00, whereas Gini Coefficient is from 0.307 to 0.313. In sum, it is for sure that the problem of income underreporting in Taiwan does exist, and the distribution of income becomes more unequal after evaluating the adjusted income.

並列關鍵字

Income Underreporting、The Survey of Family Income and Expenditure of Taiwan、The Expenditure-Based Approach、The Ratio of Household Income - Top 20% to Lowest 20%、Gini Coefficients

參考文獻

林宇文(2005)，貧窮議題法規範之研究，國立中山大學中山學術研究所碩士論 文。

李宜、鄭保志(2010)，臺灣政府各項移轉收支的重分配效果比較：1976~2006之全面性與局部性分析，經濟論文叢刊，38(2)，233-288。

連賢明(2011)，如何使用健保資料推估社經變數，人文及社會科學集刊，23(3) ，371-398。

呂逸晨(2011)，台灣地區自僱家戶低報所得之推估，國立台北大學財政學研究所碩士論文。

Benedek, D. and O. Lelkes (2009) ’’The Distributional Implications of Income Underreporting in Hungary,’’ University Library of Munich, Germany in its series MPRA Paper, 17287.