The paper develops a comprehensive framework for market risk stress testing in internationally active financial institutions. We begin by defining the scope and type of the stress test and explaining how to select risk factors and the stress time horizon. We then address challenges related to data gathering, followed by in-depth discussion of techniques for developing realistic shock scenarios. Next the process of shock application to a particular portfolio is described, followed by determination of portfolio profit and loss. We conclude by briefly discussing the issue of assigning probability to stress scenarios. We illustrate the framework by considering the development of a ‘worst case’ scenario using global financial market data from Thomson Reuters Datastream.