| Summary: | 碩士 === 國立臺灣大學 === 土木工程學研究所 === 98 === In this study, we use wavelet analysis to characterize the complex interactions between housing market and economic variables in different time-frequency domains. Different from traditional viewpoints, this paper discusses time-varying lead/lag relations in the specific frequency bands. Our study shows that their persistent interactions exist in specific frequency bands like short-term Kitchin cycle and mid-term Juglar cycle across most countries and regions. This finding reveals that these specific frequency bands have majorly cyclic features with lead/lag relations.
Variables like GDP, long-term interest rate, term spread have more consistent interactions with housing; however, policy variables like short-term interest rate and M1 are more difficult to find consistent interactions with housing across countries. Under these circumstances, the lead/lag relations of their persistent interactions still vary with time which arises from critical events or policies happening.
According to our findings, housing and GDP vary their relations of lead, lag or synchrony with time. Short-term interest rate trends to track housing only in UK and New Zealand. Long-term interest rate is a lead indicator of housing across most countries. Term spread always leads GDP and housing across most countries. Three features appear between M1 and housing: M1 is a countercyclical factor which lags Housing in US; M1 seems to follow business cycle and housing in UK; M1 leads housing in Canada and Hong Kong.
|
|---|
