Extensions to a quantitative analysis of the degree of integration between Irish and UK financial markets

Abstract

In a recent issue of this Review, Browne and O'Connell (1978) find a close relationship between the Irish and the UK interest rates simultaneous in time with the magnitudes of the UK fluctuation "completely transmitted to the Irish rate". The authors' methods are well adapted to the discovery of lags (UK rates the earlier, of course) if they exist. A typical conclusion is that the 3-month inter-bank rates are correlated only simultaneously. As regards deposit rates, "If we consider all changes in the UK rates (except for one instance . . .) the average lag in the adjustment of the Irish rate to the UK is just less than five weeks". The data relate to 201 consecutive weeks during the period 1973-1977 and deal with nine types of interest rates for both Ireland and the UK or eighteen series in all. Five pairs were selected for examination, using sophisticated methods of time series analysis. In the latter sense the paper is a pioneer in Ireland... The comments that follow are based on the Browne-O'Connell computer sheets, using methods with which the writer has been associated in his researches or mere commonsense, but avoiding the more esoteric reaches investigated by the authors with which he is unfamiliar. His main objects would be (i) to detect any periodicities in the interest rate weekly time series and (ii) to speculate on the time-lag (in weeks), if any, between changes in Irish rates following those in the UK. The computer data to be used are correlograms and power density spectra (linearly related to the simpler spectral functions used in Section III) for Irish and corresponding UK rates separately, and cross-correlations for a succession of lags between Irish and UK rates ? the UK, of course, leading.