Long-Run Trends and Cycles in US House Prices

This paper analyses US nominal house prices at an annual frequency over the period from 1927 to 2022 by means of a very general time series model. This includes both a (linear and non-linear) deterministic and a stochastic component, with the latter allowing for fractional orders of integration at b...

ver descrição completa

Detalhes bibliográficos
Autores: Caporale, Guglielmo Maria, Gil-Alana, Luis Alberiko
Tipo de documento: artigo
Data de publicação:2025
País:España
Recursos:Universidad de Málaga
Repositório:DDFV. Repositorio Institucional de la Universidad Francisco de Vitoria
Idioma:inglês
OAI Identifier:oai:ddfv.ufv.es:10641/6951
Acesso em linha:https://hdl.handle.net/10641/6951
Access Level:Acceso aberto
Palavra-chave:Cycles
Fractional integration
Long memory
Persistence
Trends
US house prices
Economics, Econometrics and Finance (miscellaneous)
Computer Science Applications
Yes
yes
Descrição
Resumo:This paper analyses US nominal house prices at an annual frequency over the period from 1927 to 2022 by means of a very general time series model. This includes both a (linear and non-linear) deterministic and a stochastic component, with the latter allowing for fractional orders of integration at both the long-run and the cyclical frequencies. The results are heterogeneous depending on the model specification and on whether or not the series have been logged. Specifically, a linear model appears to be more appropriate for the logged data whilst a non-linear one appears to be a better fit for the original ones. Further, the order of integration at the zero or long-run frequency is much higher than at the cyclical one. The former is in fact around 1 in all specified models, which implies a high degree of persistence of this component. Finally, the order of integration of the cyclical structure implies that cycles have a periodicity of about 8 years, but it is almost insignificant in all cases.