The nonstationarity of the consumption-income ratio: Evidence from more powerful Dickey-Fuller tests

Abstract

In this letter the question of whether the consumption-income ratio is mean reverting is revisited. To address known problems of low power associated with previous studies, more powerful modifications of the Dickey-Fuller (DF) test are applied. The results of weighted symmetric and recursively mean-adjusted DF tests provide strong evidence of the UK consumption-income ratio being nonstationary. This finding is further supported by reduced bias estimation of the autoregressive parameter, the estimated value of ρ derived being much closer to unity than that obtained under a conventional DF test.