L’objectif de la thèse est d’identifier la pertinence des modèles en équilibre général calculable (MEGC) pour analyser la problématique posée par les Accords de Partenariat Économique (APE) entre l’Union européenne et les pays ACP. Une revue de la littérature est d’abord réalisée, puis un modèle en équilibre général calculable (MEGC) à dynamique intertemporelle est construit pour analyser l’impact de l’APE sur un pays spécifique : le Ghana. À partir du constat portant sur la diversité des résultats de simulations, qui dépendent essentiellement de la structure du modèle et des modes de fermeture choisis par le modélisateur, ce travail cherche à mettre en évidence la largeur du faisceau de résultats possibles et l’impossibilité de mettre en avant les bénéfices potentiels qui peuvent être attendus d’un tel accord en s’appuyant simplement sur des MEGC néoclassiques standards. / This work aims to analyze to what extent the use of an applied general equilibrium model (AGE) allows to correctly assess the potential economic impact of EPAs between ACP countries and the European Union. First, a review of the literature is conducted and then an intertemporal dynamic AGE model is built in order to assess the potential impact of EPA on a specific country: Ghana. From the variety of results resulting from the models simulations and depending on hypothesis made on the model structure and the type of closure chosen by the modeler, our work aims to stress the risk of using standard neoclassical Walrasian models to assess the potential benefits of an EPA on ACP countries economy.

本研究利用台灣2001年資料以Ramsey與OLG假設分別建立區分四部門別與八部門別之動態可計算一般均衡模型，進行「提高間接稅（消費稅、環境稅）、降低直接稅（資本、勞動所得稅）」之稅制改革模擬分析。結果發現若以單一時點比較靜態方式分析可能迥異於動態長期分析結果；其次以Ramsey分析所欠缺的代際分配影響結果在OLG分析中可以察見，特別是進行以消費稅或環境稅替代資本所得稅模擬時；再者OLG因代際傳遞效果可能較Ramsey假設下反應緩慢，故也可觀察到Ramsey模型面對衝擊後若干變數增減幅度會較OLG模型陡峭。最後特別在環境稅替代所得稅模擬中經由替代稅率變化發現，活在初期世代雖可享受所得稅下降好處但卻也面臨排放稅課徵，因此福利水準增減變化不大，惟屆至排放稅課徵瀕至臨界點需使用原所得稅來平衡稅收時，此為另一波必須為減量政策付出代價的時點與世代。 / Based on a thorough comparison of the basic assumptions and economic underpinning of the Ramsey and Overlapping generation (OLG) model, this study builds both static and dynamic computable general equilibrium tax policy models to analyze the effect of tax reform on Taiwan’s economy. The new tax structure considers mainly a replacement of some parts of direct tax - labor and capital income taxes by indirect tax- consumption tax or environment tax. The benchmark year of the model is 2001, and the economy is divided into four and eight sectors. The modeling framework is implemented on top of MPSGE, which is a special module within the General Algebraic Modeling System (GAMS).Our simulation results show differential outcomes for static and dynamic models. The slower speed of transition inherent in OLG model has caused variables such as welfare, capital stock, and replacement consumption tax rate change less than that in the Ramsey model when the tax reform policy is imposed. And the intergeneration effects of different tax burden between young and old generations can be found in the OLG model but not in the Ramsey model. Finally, we find as the income tax rate coming lower and the environment tax going up when simulation of tax replacement, causing the welfare does not change much at the beginning of shock, and generations especially born in the near future when environment tax does not grow and revenue neutrality needs to be meet are another cost payers of abatement policy.