Structural Analysis of Vector Error Correction Models with Exogenous I(1) Variables

This paper presents two generalisations of the existing cointegration analysis literature. Firstly, the problem of efficient estimation of vector error correction models containing I(1) exogenous variables is considered and the asymptotic distributions of the log-likelihood ratio statistics for testing cointegrating rank are derived under different intercept and trend specifications and the respective critical values are tabulated. Tests of the co-trending hypothesis are also developed together with model mis-specification tests. Secondly, the paper considers the problem of efficient estimation of vector error correction models when the lag lengths of the included stationary variables may differ within and between equations. The purchasing power parity and the uncovered interest rate parity hypotheses are re-examined using UK data under the maintained assumption of exogenously given foreign prices.