이중차분법을 활용한 전주권역 주택시장 규제 영향 분석 = A Study on the Impact of the Housing Market Regulation in the Jeonju Region Using the Difference-in-Difference Method

다국어 초록 (Multilingual Abstract)

This research examines how regulatory policies affected the housing market in Jeonju City. After the real estate measures announced on 17 December 2020, Jeonju City became the only adjustment target area designated in Jeollabuk-do Province. The study established Jeonju City as the treatment group, with Gunsan City, Iksan City, and Wanju County serving as control groups since these neighboring areas remained unregulated. This configuration resulted in four analytical models.
Apartment transaction prices served as the dependent variable, measured as price per exclusive area and transformed using natural logarithms. To address pre-treatment differences between groups and improve estimation reliability, the research utilized Propensity Score Matching techniques. Both conventional Difference-in-Differences analysis and Sun & Abraham's enhanced approach were implemented to examine dynamic policy effects and validate the parallel trend assumption.
The DID analysis yielded varying policy effects based on control group selection. When comparing Jeonju with Gunsan, the results indicated a statistically significant regulatory suppression effect, showing apartment prices in Jeonju decreased by approximately 7.8% relative to Gunsan following the adjustment area designation. However, the Jeonju-Iksan comparison produced contrasting results with a 7.9% price increase, though this model failed to satisfy the parallel trend assumption due to significant positive effects during the pre-policy period, complicating interpretation of pure policy effects.
The Sun & Abraham analysis, which accounts for time-varying policy effects, revealed different findings. The Jeonju-Gunsan model met the parallel trend requirement and showed significant price elevation effects immediately after the adjustment area designation. This model exhibited dynamic patterns with the peak effect (approximately 9.7%) occurring four months post-designation. Given that conventional DID estimates may produce misleading results when policy effects vary temporally, the study prioritized findings from the Sun & Abraham methodology. This analysis suggests that the regulatory measures actually contributed to apartment price increases in the Jeonju area.
To verify statistical robustness, placebo tests were conducted through 20,000 random treatment reassignments. The observed effects in both Jeonju-Gunsan and Jeonju-Iksan models corresponded to distribution extremes, confirming that the policy effects were statistically significant rather than coincidental.

목차 (Table of Contents)

• 제Ⅰ장 서론 1
• 제1절 연구의 배경 및 목적 1
• 제2절 연구의 방법 및 범위 3
• 제Ⅱ장 이론적 배경 8
• 제1절 부동산 규제지역 8
• 제2절 선행연구 고찰 12
• 제3절 선행연구와의 차별성 21
• 제Ⅲ장 연구설계 및 방법론 22
• 제1절 이중차분법과 Sun&Abraham 방식의 이중차분법 22
• 제2절 연구설계 및 변수 설정 29
• 제3절 자료 수집 및 분석 35
• 제Ⅳ장 실증분석 결과 37
• 제1절 성향점수매칭(Propensity Score Matching) 37
• 제2절 이중차분법 분석 41
• 제3절 강건성 검증 55
• 제Ⅴ장 결론 62
• 제1절 연구 요약 및 주요 결과 62
• 제2절 논의 66
• 제3절 연구의 시사점 및 정책적 제언 68
• 제4절 연구의 한계 및 향후 연구 방향 71