體型對台灣全職薪資的影響：追蹤分量迴歸分析

The Impact of Body Shape on Full-Time Wage in Taiwan-A Panel Quantile Regression analyze

蔡依玲(Yi-Ling Tsai)

指導教授 ： 胡登淵

淡江大學/商管學院/產業經濟學系碩士班/碩士(2016年)

https://doi.org/10.6846/TKU.2016.00228

摘要

本文運用追蹤分量迴歸模型(Quantile Regression for Panel Data, QRPD)探討台灣男女老少的BMI與肥胖與否對於薪資之影響，資料取自中央研究院建構之「華人家庭動態資料庫」，資料期間為2004年到2011年。
本研究之追蹤分量迴歸模型除了考慮全職就業是否造成的樣本選擇偏誤性，同時也考慮BMI與肥胖是否內生，分別運用Inverse Mill’s Ratio和工具變數處理。選取的工具變數包含出生區域之肥胖、過重以上盛行率與工作區域之肥胖、過重以上盛行率。本文將男女依年齡分類，分為25-40歲為年輕族群與41-65歲為老年族群，分別各自算出生區域之肥胖、過重以上盛行率、工作區域之肥胖、過重以上盛行率。
分量迴歸的結果發現：不論是以出生區域還是工作區域作為工具變數，年輕女性之BMI對薪資影響在低分量與高分量存在負向顯著，這符合Morris (2007)的結果，BMI越高的女性越會負向影響薪資；年輕男性之BMI對薪資影響在各分量顯示為正向顯著，Cawley (2004)指出女性之BMI對薪資的影響比男性明確，但台灣實證研究結果顯示，年輕男性之BMI也是會影響薪資；老年女性之BMI則是在各分量皆為為正向顯著，可能是因為過去之價值觀與現在有差別或是老年女性可能已升為幕後主管，所以對薪資的負向影響未呈現出來；老年男性之BMI各分量皆為正向顯著。年輕女性之肥胖對薪資影響在低分量時為負向顯著，在高分量時為正向顯著；年輕男性之肥胖對薪資影響低分量為負向顯著，在高分量時正向顯著；老年男性之肥胖對薪資影響在各分量都為正向顯著，其中老年女性之肥胖無法使用現有工具變數。

關鍵字

體型 ； 薪資 ； 分量迴歸 ； 追蹤資料

並列摘要

This research uses quantile regression for panel data model (QRPD) to estimate the effect Taiwanese’s BMI or obesity on full-time wage by gender and age cohort. The data mainly come from 2004 to 2011, the Panel Study of Family Dynamics of Taiwan.
This thesis analyzes data divided into two participants’ ages: 25-40 years-old and 41-65 years-old. It uses Inverse Mill’s Ratio (IMR) to deal with samples of selection bias and the approach of instrumental variables to deal with endogeneity problem. The instrumental variables include prevalence rate of over weight in birth area , prevalence rate of obesity in birth area , prevalence rate of overweight in working area and prevalence rate of obesity in working area.
The results indicate that in quantile regression, no matter birth area or working area as instrumental variables in BMI and wage for young women, the relation between BMI and wages are significantly negative among lower qantiles and higher qantiles. It corresponds to the research of Morris’s study (2007): the higher the BMI is, the more women will negatively affect salaries; for young men the BMI positively significantly affect wage among most quantiles. Cawley (2004) pointed out that in terms of wages, for BMI, women’s are affected more than men’ s , but our empirical research in Taiwan shows that BMI also positively affects young men's wages among most of the quantiles. The older women’s BMI is positively significant among most quantiles. Those women who have more experienced may be promoted to manager working behind the scenes, so the negative impact of the salary does not appear. For old men, the relation BMI and wage is positively significant among most of the quantiles.For young women and men, the relation between fat and wages are significantly negative among lower qantiles and are significantly positively among higher qantiles. For old men, the relation fat and wage is positively significant among most of the quantiles. Obesity old women can't use the existing instrumental variable.

並列關鍵字

BMI ； wage ； quantile regression ； panel data

參考文獻

6.許雅棠"相對體型對薪資的影響：台灣老少群體的半母數分析." 淡江大學產業經濟學系碩士班學位論文 (2011): 1-67

3.郭姿伶. "台灣教育報酬率之實證研究." 淡江大學經濟學系碩士班學位論文(2008): 1-50.

4.徐美, 陳明郎,方俊德. "台灣產業結構變遷和性別歧視對男女薪資溢酬變動趨勢之影響." 經濟論文 34.4 (2006): 505-539.

23.Shimokawa, Satoru, Hung-Hao Chang, and Per Pinstrup-Andersen. "Understanding the differences in obesity among working adults between Taiwan and China." Asia Pacific journal of clinical nutrition 18.1 (2009): 88.

24.Shimokawa, Satoru, Hung-Hao Chang, and Per Pinstrup-Andersen. "Understanding the differences in obesity among working adults between Taiwan and China." Asia Pacific journal of clinical nutrition 18.1 (2009): 88.