This paper provides an approach to testing the measures of the one-way effect for cointegrated vector time-series in the presence of trend breaks. We propose Wald testing the measures and their computational algorithm, an extension of previous work by Hosoya and Yao and Hosoya, to the case where trend breaks are explicitly taken into account in the cointegration relationship. On the basis of the proposed inferential method and the derived evidence, we present a causal structure characterization of money supply and income as well as interest rates for the last 44 years of the Japanese economy, and contrast it with the results of Yao and Hosoya.