Dimensional affect prediction from speech has traditionally relied on acoustic features to estimate continuous affect representations (e.g., arousal, valence) at each time step. However, affect evolves dynamically over time, and incorporating temporal information may improve prediction accuracy. This study investigates emotional dynamics in speech emotion recognition using multitask learning, where a model jointly predicts both the affect state and its temporal derivative. Experiments on the RECOLA and SEWA datasets show that incorporating dynamic information improves affect state prediction, particularly for valence, known to be challenging to model from audio alone. While CCC scores for affect dynamic predictions remain lower than those for affect state predictions, results indicate that learning dynamics as an auxiliary task enhances affect state estimation over time. These findings underscore the importance of modelling emotional dynamics to capture the temporal evolution of affect.