台灣不同世代之終身薪資軌跡

THE LIFETIME EARNINGS ACROSS COHORTS IN TAIWAN

本文使用「華人家庭動態」資料庫之1999－2018年追縱資料，並利用混合模型（mixed model）之隨機效果，將個人薪資偏離母體薪資軌跡的幅度分為兩部分，一為因勞動者的初始條件因素，使薪資永久偏離。二為個人薪資成長的差異，使薪資偏離隨年齡而變化。實證發現初始條件薪酬隨世代而下降，1981－1990世代的男性，若擁有高中、專技、大學等學歷，相對於1960前世代相同學歷的初始條件薪酬，分別減少了34％、32％、50％。在這20年的樣本期間，並無證據顯示出生世代影響薪資成長；而是個人的初始條件超額薪酬（或折價），將使終身薪資軌跡產生差異。不管是年長或年輕世代，個人有較佳（較差）之未觀察到能力，不僅使初始條件薪酬相對偏高（偏低），同時有較高（低）的終身薪資成長。

The study adopts the Panel Study of Family Dynamics database which comprises 20-year longitudinal surveys to examine cohort differences by estimating the heterogeneous lifetime earnings among individuals. The random effects of the mixed model are used to distinguish which portion of lifetime earnings was due to the remuneration of initial conditions and which part was due to age-earning profile. Empirical results show that the remuneration for initial conditions declined over cohorts. If men born in the 1981-1990 cohort possess high school, college, or university education, respectively, then their remuneration for initial conditions compared with men born before the 1960 cohort will be reduced by 34%, 32%, and 50%, respectively. In the past 20-year surveyed period, no evidence shows that an individual's cohort would cause earnings to grow heterogeneously. Regardless of any cohort, an individual's extra (or discounted) remuneration for initial conditions would affect the growth of lifetime earnings. Those with superior (or inferior) unobserved abilities would have higher (or lower) remuneration for initial conditions, which also increases (or reduces) the growth of lifetime earnings.