本研究論文採用空間分量迴歸模型來探討所得不均與國家生產效率等兩個經濟議題，該模型同時考慮空間外溢性及異質性之特性，可有效避免推估的偏誤。本文內容架構分為五章，第壹、貳章為前言與本文所使用之計量模型與資料包絡法之介紹，接續兩章為兩篇實證研究，最後一章為結論。 第參章題名為：「數位落差與所得不均」，文中採用空間分量迴歸模型探討所得不均的空間相關性及其影響因素。我們以基尼係數表示所得不均，以網際網路使用率表示數位落差。首先，我們分析全球數據結果發現，所得不均存在空間相關性，數位落差對其有顯著的正向影響。其次，我們將全部資料依高所得和低所得國家分群，兩群估計結果迥然不同。高所得國家中，所得不均仍存在正向之空間相關性，在低所得國家則為負向的，此外，網際網路使用率越高在高所得國家可降低所得不均，反之在低所得國家則將加劇所得分配不均，另增加中學入學率可以縮小所得差距，特別是在低所得國家。 第肆章題名為：「政府貪腐、非意欲產出與國家生產效率」。本章使用兩種資料包絡分析方法來估計國家生產效率，一為不考慮非意欲產出的生產效率，另一為考慮二氧化碳的排放量做為非意欲產出計算所得出之生產效率。我們比較兩種生產效率發現歐洲國家近年來對環境所投入之人力、物力已見成效，此與以往相關之文獻結果不同。由模型推估結果發現，若以不計非意欲產出之生產效率做為應變數時，我們發現低生產效率國家有顯著空間相關性，高生產效率國家則無；貪腐程度則對高生產效率的國家影響較大，外人直接投資對低效率國家可顯著改善所得不均，貿易依存度則可能對高生產效率的國家有負向的影響。若以考量非意欲產出之生產效率來看，發現空間相關性於部分分量仍存在，惟效率較低的國家將無法透過周邊國家外溢獲得益處；而貪腐程度對高生產效率國家之負向影響將遠大於低效率的國家，此乃因為高效率國家之貪腐程度越高將使得二氧化碳排放增加。最後我們發現提高貿易依存度反而可能將引發低效率國家之生產效率降低。 This dissertation uses the spatial quantile regression model to discuss the two economic issues such as income inequality and national production efficiency. The model can consider the spatial spillover effect and the heterogeneity of data, which can effectively avoid the bias. The dissertation is divided into five chapters, the first two chapters are the introduction and the econometric models and the data envelopment analysis method. The following two chapters are two empirical articles. The final chapter is the conclusion. The third chapter titled “Income Inequality and Digital Divide”. Firstly, we analyze the whole data. The estimation results find that income inequality is positively spatial dependent across regions and the Internet has a significantly positive effect on income inequality. Secondly, we divide entire data into two groups of countries, viz., high and low income countries. The estimation results of two groups are quite different. The income inequality were positively spatially correlated among neighboring countries in high-income countries but negatively in low-income countries. On the other hand, the Internet usage exacerbate income disparity in low-income countries but improve income inequality in high-income countries. The results also show that increasing school enrollment can alleviate income gap especially in low-income countries. The fourth chapter titled “Corruption, Undesirable Output and Efficiency”. This article uses two types of DEA(data envelopment analysis) method to estimate the production efficiency for countries: one did not consider the undesirable output and the other one considered the carbon dioxide as the undesirable output to obtain the production efficiency. By comparing the results with these two types of efficiency estimation, we found that the efforts by European countries in recent years has been paid off, which is different from the past literatures. We found that the degree of corruption affected high production efficiency countries more significant. Foreign direct investment had a significant effect on less efficient countries. Trade might have a negative impact on the highly productive countries. By considering undesirable output, while spatial correlation existed in some quantiles, the least efficient countries would not get the benefits through the spillover effect from adjacent countries. The extent of corruption influenced highly productive countries more negatively than inefficient countries, since the high degree of corruption in the higher productive countries would allow an increase in emission of carbon dioxide. Finally, we found that increase in trade will be likely to lead to lower production efficiency in less efficient countries.
