|摘要: ||過去四十多年來，許多開發中國家陸續推動貿易自由化政策，藉由這些政策，這 些國家的政策制訂者冀望能提高其國內的經濟成長以及生活水準。然而，貿易自由化 政策，也可能會影響其國內資源的分配，進而影響所得分配的結果，因此，為能全面 評估經濟自由化所帶來的經濟效應，除了經濟成長的成效外，貿易自由化可能引起的 所得分配效果理應也須謹慎的加以檢驗。一般而言，貿易自由化會透過改變相對勞動 所得，改變相對價格進而影響消費，改變個人從事生產活動決策等三個管道影響個人 所得分配。經由這三個管道，貿易自由化對所得分配的最終影響，須由實證的證據來 檢視。但利用開發中國家跨國資料所進行的貿易自由化之所得分配效果的研究，通常 並未獲得一致性的結論。而且，貿易自由化影響所得分配的機制也可能因國家而異， 而從跨國資料所得到的結論，通常是基於”平均”的研究結果，可能會因忽略個別國家 的特殊性，以致無法針對個別國家因經濟自由化所帶來的所得分配情況得到可靠的資 訊以進行合理的推論。此外，利用跨國資料進行迴歸分析所得到的推定量通常會因內 生性的問題產生偏誤的估計。為避免以上所提之問題，本研究將應用合成控制法 (Synthetic Control Method)對一組國家進行由資料驅動(data-driven)的比較案例分析，以 重新檢驗貿易自由化影響開發中國家所得分配變化的效果。此方法的優點是針對單一 受政策影響國家與其一組對應控制組的國家，進行比較分析，並藉由考慮隨時間變化 所觀察不到的干擾項，解決因遺漏變數所引起的內生性問題。透過此新進的計量方法， 吾人將對選定的開發中國家，檢驗貿易自由化對所得分配變化的影響。本研究期望能 對於貿易自由化的所得分配效果提供更進一步的實證證據。
Theoretical prediction in international economics and growth theory largely points to a positive relationship between economic liberalization and economic growth. As such, in the past four decades, many developing countries have undergone significant trade liberalization that substantially increased their exposure to international markets. By so doing, the policy makers of these developing countries are in the hope of achieving sustained growth and higher economic welfare with these liberalizing measures. Nevertheless, changes in a country’s exposure to international trade, or world markets more generally, can also affect the distribution of resources within the country and hence generate substantial distributional conflict. As such, to fully assess the economic consequence of globalization, the impact of trade liberalization on inequality should be carefully examine as well. Generally, trade liberalization may affect individuals through three main channels: changes in their labor income, change in relative prices and hence consumption, and changes in household production decisions. But the ultimate distributional effect of trade liberalization should be provided by empirical evidence. However, maybe due to different liberalization and inequality measurements, samples, periods and econometric techniques employed, empirical results from cross-countries studies generally yield inconsistent conclusions on the distributional effect of trade liberalization in developing countries. Furthermore, different countries may experience trade liberalization in different ways and at different time, so the mechanisms through which inequality is affected may be case specific. As such, cross-country studies that draw conclusion on average of the data usually ignore the heterogeneous experiences and circumstances faced by different countries, and hence generate no reliable inference for individual country. In addition, the cross-countries estimators usually suffer the endogeneity problem and thus likely result in biased estimations. To abstain from the problems mentioned in the previous studies, I will revisit the issue of distributional effect of trade liberalization by implementing a set of empirical country case studies. In particular, I will apply a recently developed econometric technique, the synthetic control method (SCM, hereafter), to implement data-driven comparative case studies. the advantage of the proposed SCM framework is that it can deal with endogeneity from omitted variable bias by accounting for the presence of time-varying unobservable confounders. This feature improves on panel data models such as fixed effects or difference-in-difference, which can account for only time-invariant unobservable confounders. Accordingly, this approach can be viewed as an alternative for standard cross-countries estimators that are prone to suffer endogeneity issues. In this manner, this research hopefully can complement the extant literature by delivering more country specific information of distributional effect resulting from trade liberalization for some developing countries and thus can further provide more insights and policy implications from the analyzed results.