勞動生產力與聚集經濟──來自空間相關誤差組成縱橫模型之證據

Labour Productivity and Agglomeration Economy - Evidence from Panel Models with Spatially Correlated Error Components

本研究估計日本47個都道府縣之聚集經濟效果，運用晚近新發展之空間相關誤差組成縱橫模型與一個先前未被分析過的社經統計資料組橫跨47個空間單元從1990年到2012年，並考量到相依變數與自變數之間的內生性議題，在控制勞動效率變異的情形下，實證結果為當就業密度增加一倍時，勞動生產力會增加4.52%，顯示出聚集經濟的確對於勞動生產力有顯著的影響。

The research estimates the effects of agglomeration economy for 47 Japanese prefectures. We adopt a newly developed panel model with spatially correlated error components and use a socioeconomic dataset containing data across 47 spatial units from 1990 to 2012 and not previously analysed. The endogeneity issue between dependent and independent variables is also properly considered. While controlling for the variation of labour efficiency, the empirical results show that a doubling of employment density increases labour productivity by 4.52%, and agglomeration economy does play a significant role in boosting labour productivity.