本文主要之研究目的,是以工作/不工作(WORK)以及薪資(LNWAGE)兩個面向觀察台灣地區本籍配偶與外籍配偶就業行為決策之影響因素以及比較兩組群之間是否存在差異性。本研究選用2003年「台灣地區婦女婚育與就業調查」、「外籍與大陸配偶生活狀況調查」以及2008年「台灣地區人力運用調查」、「外籍與大陸配偶生活需求調查」兩年度的原始資料進行實證分析。在計量分析上,本研究採用probit模型估計就業方程式,以及採用OLS模型估計薪資方程式;此外,亦探討本籍配偶與外籍配偶就業行為決策之差異。吾人利用Yun(2004)之非線性分解方法(probit decomposition)估計就業差異模型,採薪資分解作法(Blinder-Oaxaca decomposition technique)估計薪資差異模型。由於外籍配偶2003年與2008年這兩年度間就業率大幅上升,因此本文亦對於外籍配偶就業差異行為之轉變進行實證分析。 就業方程式實證結果發現,本籍配偶較受自身之人力資本所影響,家庭背景因素影響力較小;而外籍配偶除受本身歸化程度影響以外,也受家庭背景因素影響。就業差異模型之實證結果為,模型中係數差異佔總就業差異相當大的比重且為正向影響;其中「年齡」、「移民年數」以及各自之平方項之加總效果最為明顯。薪資方程式實證結果顯示,本籍與外籍配偶之薪資都較受「個人特徵變數」以及「工作特徵變數」影響。薪資差異模型之特徵差異減少了兩者之間的薪資差異,而係數差異則增加了兩者之薪資差異,其中「移民年數」及「移民年數之平方項」顯著度較高。 最後,由2003年與2008年實證結果之比較得出,不論是東南亞籍配偶或是大陸籍配偶,就業率均大幅度上升,而此就業行為上轉變的主要原因是由個人特徵背景的變化所致。
The purpose of this thesis is to examine employment status between foreign spouses and native-born spouses in Taiwan. We estimate the employment probability differential based on Yun’s non-linear decomposition model. The wage differential analysis is based on Blinder-Oaxaca decomposition technique. The empirical analysis has been carried out using data from both The 2003 and 2008 Survey of Foreign and Mainland Spouses’ Lift status collected by Ministry of Interior, R.O.C., The 2003 Survey of Women’s Marriage, Fertility and Employment as well as The 2008 Manpower Utilization Survey collected by National Statistics, R.O.C. Employment decision of native-born spouses is strongly correlated with her human capital, but slightly correlated with their family background. However, employment decision of foreign spouses is strongly correlated with their citizenship status, also strongly correlated with their family background. The employment differential between foreign spouses and native-born spouses is almost entirely due to difference in coefficient, especially in the coefficient of age, year since migration and their squared terms. Wage of native-born spouses and foreign spouses is strongly correlated with their individual characteristic variables and job characteristic variables. It is found that the difference in characteristic reduces the wage differential between the two groups, but the difference in coefficient increases their wage differentials.