本研究的目的是評估不同勞動所得租稅抵減制度(Earned Income Tax Credit, EITC) 設計對勞動供給之影響。本文以雙元誤差模型 (two-error model, TEM) 處理勞動者面臨之分段線性預算限制式, 利用95年主計處人力運用調查資料輔以95年家庭收支調查資料設 算非勞動所得, 分別估計20至60歲已婚夫婦的結構式勞動供給函 數。接著模擬三種不同EITC 方案下, 分析家戶勞動供給的變化。 本文實證結果發現使用最小平方法或Tobit 與TEM 模型的估計 結果明顯不同, 逕以縮減式估計勞動供給函數將可能產生內生性問 題。其次, 妻子工時聚集於折點處的現象較丈夫明顯, 顯示妻子有 較大的替代彈性。第三, 在不同的EITC 設計下, 模擬分析發現隨 著最高抵減額的增加, 勞動工時選擇出現往該區段聚集的現象, 且 區段選擇跳動情況明顯。以最高抵減額48,000元,遞增遞減率各為 40%與20%為例, 將使丈夫、妻子平均分別減少6.9%、8.4%的工 作時數。最後, 不論最高抵減額大小, 平均將使59%的無業家庭投 入就業市場, 雖然工作時數的增加有限。

Nowadays Earned Income Tax Credit (EITC) has become one of the most expensive and fastest growing social welfare programs in America. As a result of the success of the EITC in America, this program has also been successfully instituted in many other countries. This paper evaluates the potential effect of an EITC on labor supply in Taiwan. The traditional econometric methods estimating reduced form labor supply by a Tobit or OLS will lead to serious endogeneity problems, when labor faces a kink-point-piecewiselinear budget. We apply the Two Error Model (TEM) in Triest (1990) to overcome this issue. Our finding reveals first that wives’ labor hours tend to cluster more at the kink point than husbands do, and the elasticity of substitution for wives is larger than that for husbands. Secondly, the simulation of an EITC shows that when the largest tax credit rises, workers’ labor hours are likely to be in the largest tax credit region. When the largest tax credit is 48,000, phase-in and phase-out rates are 40% and 20% respectively. An EITC program could cause a husband and wife on average to reduce their labor hours 6.9% and 8.4% respectively. Finally, no matter what the largest tax credit is, on average, the EITC program will encourage about 59% of non-working family to work.