以受限概念探討家庭生育決策

A Study Of Household Fertility Decision: Application Of Censored Regression Model

黃宜葳(I-Wei Huang)

指導教授 ： 李鈞元

淡江大學/商管學院/經濟學系碩士班/碩士(2014年)

https://doi.org/10.6846/TKU.2014.00086

摘要

由於現今女性教育程度提升、勞動參與增加與傳統觀念改變；因此，各國生育率皆日趨下降，進而引發對人口老化的憂慮。所以，本文利用行政院主計總處2010年「婦女婚育與就業調查」的資料，以受限的卜瓦松分配來探討影響已婚婦女生育決策的因素；實證結果指出：各解釋變數的正負，包含婦女、配偶個人特質變數的正負，無論在一階段迴歸或兩階段迴歸大多符合預期。與既有研究比較，多數文獻探討生育決策時並未加入受限概念，自然必須捨棄未完成生育歷程的資料(約以50歲為界)；而本文修正此項缺點以保留樣本全貌，樣本數的擴大不僅使得迴歸估計更有效率，並且可觀察不同世代婦女在生育行為上的變化。

關鍵字

生育 ； 受限 ； 卜瓦松分配 ； 兩階段迴歸

並列摘要

Due to the enhancement of female education, to the increase of the female labor participation, and to the change of female concept, there are the decline of fertility rate and then the aging of population in several countries. Hence, this paper is going to use the Censored Poisson Model to discuss the fertility decisions of married women. Our data are from the Taiwan's “Women’s Marriage, Fertility and Employment Survey” in 2010, and the empirical results show that the signs of many explanatory variables, including woman’s and her husband’s personal characteristics variables, conform with usual expectation either in one-stage model or in two-stage model. Comparing to given researches, most of articles don't add the concept of censoring in their model, and then have to abandon those data which haven't completed the process of fertility. However, this paper correct these shortcomings to keep the sample complete, and then can improve the regression's efficiency and can observe the behavior's change from different cohorts.

並列關鍵字

fertility ； censored ； Poisson distribution ； two-stage regression

參考文獻

王德睦、張國偉 (2010) “台灣女性初婚率的量與步調”，臺灣社會福利學會，

楊靜利 (1996) “婦女勞動參與對生育率的影響”，臺灣社會學刊，19，35-56。

駱明慶 (2007) “台灣總生育率下降的表象與實際”，研究台灣，3，37-60。

林彥伶 (2013) “老年經濟保障對生育的排擠效果－台灣的實證分析”，中央研究院經濟研究所經濟論文，41(2)，255-295。

陳文意、周美伶、林玉惠、陳明吉 (2013) “抑制房價以提高生育率：以台北都會