This paper performs a multivariate cointegration analysis of UK money demand 1873-2001, and illustrates how a long-run time series analysis may be conducted on a data set characterized by turbulent episodes and institutional changes. We suggest accounting for the effects of the two world wars by estimating additive data corrections, thereby allowing the propagation of war-time shocks to be fundamentally different from the transmission of peace-time innovations. In addition, the corrected series may be used in counterfactual event studies to assess the impacts of special events. In the empirical analysis we find a single equilibrium relationship relating velocity to opportunity costs, and we identify a significant link between excess money and inflation. After accounting for the turbulent periods, the equilibrium structure is reasonably stable over a period of 130 years