Both theory and simulation have played important roles in defining and illuminating the key mechanisms involved in substorms. Basic theories of magnetic reconnection and of interchange and ballooning instabilities were developed more than 50 years ago, and these plasma physical concepts have been central in discussions of substorm physics. A vast amount of research on reconnection, including both theoretical and computational studies, has helped provide a picture of how reconnection operates in the collisionless environment of the magnetosphere. Still, however, we do not fully understand how key microscale processes and large-scale dynamics work together to determine the location and rate of reconnection. While in the last twenty years, it has become clear that interchange processes are important for transporting plasma through the plasma sheet in the form of bursty bulk flows and substorm expansions, we still have not reached the point where simulations are able to realistically and defensibly represent all of the important aspects of the phenomenon. More than two decades ago it was suggested that the ballooning instability, the basic theory for which dates from the 1950s, may play an important role in substorms. Now the majority of experts agree that regions of the plasma sheet are often linearly unstable to ideal-MHD ballooning. However, it is also clear that kinetic effects introduce important modifications to the MHD stability criterion. It is still uncertain whether ballooning plays a leading role in substorms or has just a minor part. Among the different types of simulations that have been applied to the substorm problem, global MHD codes are unique in that, in a sense,they represent the entire global substorm phenomenon, including coupling to the solar wind and ionosphere,and the important mechanisms of reconnection, interchange, and ballooning. However, they have not yet progressed to the point where they can accurately represent the whole phenomenon, because grid-resolution problems limit the accuracy with which they can solve the equations of ideal MHD and the coupling to the ionosphere, and they cannot accurately represent small-scale processes that violate ideal MHD.