This paper analyzes the interaction between exchanges trading in identical assets. Issues like price leadership, market spreads and activity/volume are related to different trading systems. Bid-ask spread estimation is conducted for each market individually taking account of conditional expectations. A VECM-GARCH (vector error correction with generalized autoregressive conditional heteroscedasticity) model incorporates the modeling implications of these findings when extending the analysis to a multivariate setting. Both univariate and multivariate tools are applied to the competition in BUNts futures trading between LIFFE (London International Financial Futures Exchange) and DTB (Deutsche Terminborse). At the same time, a computerized dealer system (DTB) is compared to an open outcry system (LIFFE). In a broader context, this paper therefore has implications for the survival potential of duplicative contracts traded at simultaneous markets under different trading systems.