團結真有力，阻礙真有心：探析拉丁美洲不平等、二元化與重分配支出效果

Abstract

這是一本探討經濟不平等與勞工階級的內部分歧如何共同作用，來影響拉丁美洲國家再分配支出的論文。論文的發問起源於這樣的觀察：拉丁美洲國家長久以來的高度經濟不平等，以及即便拉丁美洲國家在勞工的力量以及社會福利支出都比其他發展中國家來得高，為何仍然無法有效降低經濟不平等的問題。從既有研究來看，這個現象尚無法被比較福利國家文獻中的權力資源理論有效解釋，也與戰後西歐福利國家擴張的歷史經驗並不一致。我的論點起始於大多數發展中國家的現實特徵：大規模的非正規部門勞工。我透過借鑒David Rueda的「局內人－局外人」理論和Meltzer & Richard模型，提出本文主要論點：經濟不平等的惡化，將使得正規部門的工人傾向於阻礙擴大再分配政策。正規部門和非正規部門工人之間的分歧，意味著對再分配政策的不同偏好，而經濟不平等會加劇來自非正規部門勞工的再分配需求。在這樣的條件下，正規部門勞工可能會阻礙進一步的再分配，因為他們會是再分配支出擴張的主要稅收承擔者。 在實證部分，本文主要分為三個部份來驗證核心論點。本文首先使用涵蓋1985年至2002年間，15個拉丁美洲國家的跨時橫截面數據來檢驗論點，並對統計結果進行穩健性檢測，去看研究發現是否會因為改變主要的自變量、樣本範圍條件，以及模型設定而有所不同。此外，本文也檢驗了主要論點所預期的可觀察意涵，試圖評估本文論點在分解再分配支出後，以及在考慮非正規部門大小對論點的影響後，是否仍然有效。最後，本文從二手文獻資料中，援引巴西、阿根廷與智利等三個簡短個案來評估因果機制，並使用量化分析不平等如何動員勞動抗議，以及勞工動員又如何進一步阻礙再分配。整體而言，本論文有助於增進我們理解勞工階級分化會如何影響發展中國家的再分配情勢，並提出一個新觀點來解釋發展中國家的再分配過程。

Why do so many Latin American countries fail to address economic disparities, even though the political power of organized labor and the level of social spending in this region is stronger and more generous than its counterparts in the rest of the developing world? My argument starts with the reality characterizing most of developing countries: the large size of the informal sector. Drawing on insider-outsider theory and the Meltzer and Richard model, I argue that formal-sector workers will tend to impede expanding redistributive policies as economic inequality worsens. Labor divide arising from the segmentation between formal-sector and informal-sector workers indicates different preferences for redistributive policies. Moreover, increasing inequality intensifies the conflicting interests because of more redistribution demand from informal-sector workers. In response to the redistributive pressure, labor market insiders may hinder further redistribution because they may suffer from more taxation that is used to finance more redistributive spending. The empirical tests consist of three parts. I first test this argument using time-series-cross-section data covering 15 Latin American countries from 1985 to 2002. The empirical results are robust to a variety of settings, including changing the main independent variable, scope conditions, and model specifications. Furthermore, I also test the observable implications that derive from my argument, seeking to assess if my theory still works after disaggregating redistributive spending and after accounting for the role informal sector plays in moderating my theoretical expectation. Finally, I evaluate causal mechanisms underlying correlation analysis in the previous results by providing three abbreviated cases from secondary sources, as well as quantitative analysis on how inequality mobilizes labor protests and on how labor mobilization, in turn, impedes redistribution. Overall, this thesis contributes to the emerging literature on the impact of working class divide on redistribution in the developing world.