營建投資與經濟成長之關聯性-以VAR與小波轉換分析

Abstract

本研究使用VAR、Granger因果關係檢定與小波轉換分析台灣與美國營建投資與經濟成長之關聯性。根據行政院主計處編製之國民所得統計年報定義，營建投資包括：住宅、非住宅房屋(如廠房、校舍等)、公共建設(如道路、機場等)。 Granger因果關係檢定結果顯示台灣營建投資與經濟成長存在回饋(feedback)關係，表示營建投資與經濟成長互相影響對方。進一步將營建投資分為公共建設與非公共建設分析(含住宅與非住宅房屋)，結果顯示GDP單向影響公共建設與非公共建設。小波轉換則顯示台灣營建投資在8-12年景氣循環週期正向領先經濟成長，但其領先幅度隨著時間逐漸減小。接著將營建投資各分項分別以小波轉換分析，結果發現台灣營建投資各分項中以非住宅房屋與經濟成長的同調性(coherence)最大，表示其與台灣過去經濟成長的相關程度最高。 美國Granger因果關係檢定則顯示營建投資單向影響GDP，公共建設與經濟成長為獨立沒有因果關係，非公共建設則單向影響經濟成長，表示營建投資對經濟成長的貢獻主要來自非公共建設。小波轉換則顯示美國經濟成長在4-8年景氣循環週期正向領先營建投資，而營建投資各分項中以住宅與經濟成長的同調性最大。

This study uses VAR、Granger causality tests and wavelets to analyze the relationship between construction investments and economic growth in Taiwan and U.S. Construction investments are defined as residential investments and other buildings (such as factory buildings and school buildings) and infrastructure (such as roads and airports). Granger causality tests show that there is a feedback relationship between construction investments and economic growth in Taiwan, and it means that construction investments and economic growth mutually influence each other. Further, construction investments are classified into infrastructure and non-infrastructure (including residential investments and other buildings), and we find that GDP Granger causes infrastructure and non-infrastructure. Wavelets show that there is a positive relationship between construction investments and economic growth on the 8-12 year frequency band with construction investments leading, but the difference between construction investments and economic growth is gradually decreasing. Further, we use wavelets to analyze the components of construction investments, and we find that other buildings are the most coherent with economic growth among the components of construction investments, and it means that other buildings have the highest degree of correlation with economic growth. In U.S., Granger causality tests show that construction investments Granger cause GDP; there is an independent relationship between infrastructure and economic growth; non-infrastructure Granger causes GDP. It means the contribution of construction investments to economic growth mainly from non-infrastructure. Wavelets show that there is a positive relationship between construction investments and economic growth on the 4-8 year frequency band after 1970 with economic growth leading, and we find that residential investments are the most coherent with economic growth among the components of construction investments.