선거와 재정정책 사이의 인과관계에 대한 시계열회귀분석 = 한국과 대만에서의 선거의 정치경제

다국어 초록 (Multilingual Abstract)

This research paper intended to identify the causal relationship between elections and incumbent governments' fiscal policy. The first hypothesis of this paper was deduced from the review of the debates on the relevance of the Political Business Cycle (PBC) model which shed light on the causal effect of elections on the economy. From this review, I discovered such necessary conditions as "the fixed timing of elections," "incumbent governments' majority status in the legislature" and "manageable economic condition," and then formulated the first hypothesis on the basis of these three conditions. This paper derived the second groups of hypotheses from the relaxation of an assumption of the PBC model with regard to the extent of electoral influence on incumbent governments' policy incentive to manipulate fiscal policy. The relaxation of the assumption drove me to focus on such new variables as "regime types" and "the proximity to election dates" which constituted the independent variables of the second and third hypothesis, respectively.
This paper constructed basic three statistical models to test three hypothesis and then conducted the autoregressive integrated moving average (ARIMA) time-series regression analysis. The ARIMA regression analysis of the first statistical model revealed that Korean incumbent governments have pursued expansionary fiscal policy one quarter before elections, while Taiwan governments have done so three quarters before elections.
The analysis of the second and third model disclosed that Korean and Taiwan incumbent governments' fiscal manipulation varied with the extent of electoral influence. However, the way that they were affected by the extent of electoral influence was different. While the probability and extent of Korean governments' fiscal manipulation changed according to the types of regime, Taiwan governments' pre-electoral fiscal policy were not subjected to the variable of regime types but rather to that of the proximity to election dates.
The research paper run the F tests to identify which one has more explanatory power between two groups of regression models. The F tests told us that the second groups of the regression models underpinning hypotheses derived from the relaxation of the assumption on electoral influence were better than the first model which was based on the assumption that the effect of elections on incumbent governments' economic policy would be the same in every election.

목차 (Table of Contents)

• Ⅰ. 서론
• Ⅱ. 가설의 구성
• Ⅲ. 가설검증의 방법론적인 문제 : 변수측정, 사례선택, 회귀모델
• Ⅳ. 회귀분석 결과 및 해석
• Ⅴ. 결론
• 참고문헌
• abstract