Affine processes are distinguished by their rich structural properties while also retaining tractability. The defining feature of affine processes is the exponential affine form of the characteristic function of its transition probabilities which can also be characterized by ordinary differential equations. This allows for efficient computations which makes affine models useful in finance.
So far affine processes have only been considered under the assumption of stochastic continuity. However, some time series in finance exhibit jumps of random height at priorly fixed dates such as after board meetings or elections. This is inconsistent with stochastic continuity. Therefore we study affine semimartingales beyond stochastic continuity. Under fairly general conditions we are able to completely characterize the conditional characteristic function of such processes in terms of an abstract measure differential equation. On the other hand we prove existence of affine semimartingales given a set of certain parameters.