當「標準差法」遇上「ORU薪資估計模型」：探討過度教育之實際分配法

Estimating ORU Extended Wage Equation When Measuring Overeducation With Realized Matches.

何友仁(Yu-Jen Ho)

指導教授 ： 湯蕙瑄

國立臺北大學/社會科學學院/經濟學系/碩士(2015年)

摘要

本文使用「台灣社會變遷基本調查」第二期第三次(1997)、第三期第三次(2002)與第六期第三次(2012)的調查資料，針對研究過度教育常用之實際分配法提出二項建議。第一、重新定義過度教育年數、適度教育年數與不足教育年數。此方法不僅能避免過度（不足）教育年數對薪資之影響被高估（低估），更可使實際分配法中ORU薪資估計模型與V-V薪資估計模型之結果能互相呼應。第二、為了配合台灣在過去幾十年迅速變遷的社會狀況，本文以「年代標準差法」取代標準差法。考慮了年代的差別後，可避免標準差法中過度教育者集中在年輕世代的問題。實證結果發現，在使用年代標準差法時，過度教育年數與不足教育年數對薪資的影響明顯下降，且過度教育者的薪資報酬低於適度教育者。除此之外，為討論實際分配法中過度教育者與適度教育者間的薪資差異，本文在薪資估計模型中加入「自我評估有無過度教育」之虛擬變數。實證結果顯示，在控制了自我評估有無過度教育後，過度教育者與適度教育者薪資報酬的差異變小，進而推論年代標準差法中，過度教育者身上有自我評估過度教育者的人格特質。

關鍵字

過度教育 ； 薪資

並列摘要

Making use of the data of 1997, 2002 and 2012 Taiwan Social Change Survey, this thesis proposes two modifications on the realized matches method in studying overeducation. The first modification redefines the years of overeducation, required education and undereducation. The proposed definition not only avoids the possibility that the returns to overeducation (undereducation) being over- (under-) estimated, but also makes it possible to compare the empirical results estimated using the ORU wage equation versus the V-V wage equation. In order to take into account the rapidly changing labor markets in Taiwan, the second modification replaces the realized matches within an occupation with the realized matches within an occupation for the same cohort. When we further restricted the realized matches within both occupation and cohort, the resulting overeducated workers no longer overly concentrate on younger persons. The empirical results show that the returns to overeducation is smaller than that reported in the literature. Yet the returns to overeducation is consistently smaller than the returns to required education.
In addition, we propose to control for whether one is self-assessed as overeducated or not in the wage equations. The empirical results show that the difference between the returns to overeducated and adequately-educated years narrows down, which implies that the overeducated workers defined using realized matches method may have some unobservable characteristics shared by those who self-assessed as overeducated.

並列關鍵字

overeducation ； wage

參考文獻

莊奕琦、賴偉文（2011b）。不同世代下之教育報酬與能力差異。經濟論文叢刊，81-114。

鄭淳仁（2010）。高等教育擴張政策對於過度和不足教育者之影響。台北大學經濟學系研究所碩士論文。

Burris, V. (1983).The social and political consequences of overeducation. American

Sociological Review, 48(4),54-67.

Chevalier, A. (2003). Measuring overeducation. Economic Review, 70, 509-531.