A Time Series Analysis of U.K. Construction and Real Estate Indices
Posted: 20 Feb 2013
Date Written: 2013
Abstract
This study assess the nonlinear behavior of U.K. construction and real estate indices. Standard unit root tests show that both time series are I(1) processes. However, the empirical results show that the returns series for both indices deviate from the null hypothesis of white noise. Moreover, we have found evidence of nonlinearity but strong evidence against chaos for the returns series. Further tests show that the source of nonlinearity is rather different. Hence, the construction index return series displays weak nonlinear forecastability, typical of nonlinear determinisitic processes, whereas the real estate index could be characterized as a stationary process about a nonlinear deterministic trend.
Keywords: nonlinearity, heteroskedasticity, random walk, Chaos, nonlinear predictability
JEL Classification: G12, G14, G15, R30
Suggested Citation: Suggested Citation