Determining what drives stock returns: Proper inference is crucial: Evidence from the UK

This paper employs a century of the UK stock market data to examine various sate-space model specifications and Vector Autoregression (VAR) models to investigate how much expected returns and expected dividend growth contribute to movements in the UK price–dividend ratio. We show that the results of the estimated state-space models and the estimated VAR return decomposition models that attempt to estimate the contribution of expected returns and dividend growth to movements in the price–dividend ratio provide different results when one corrects for proper inference for both models. The corrected inference indicates that the contribution of expected returns to fluctuations in the price–dividend ratio is found to be statistically insignificant according to the state-space model, however, expected returns are found to contribute significantly to movements in the price–dividend ratio when one employs the VAR model. We offer some important econometric insights about the reasons for why state-space models and VAR models may give different results.