Abstract
The cointegrated nature of UK consumption and income is examined using ‘official’ seasonally adjusted data and data which have been seasonally adjusted using an alternative well-established method. It is found that the presence of cointegrating relationship is dependent upon the method of adjustment used, a result reflecting the warnings of Wallis (Journal of the American Statistical Association, 63, 18–31, 1974). This result has major implications given the plethora of cointegration studies conducted using seasonally adjusted data.