We also explore the robustness of our model when second period valuations are correlated across individuals, and when each individual's first and second period types are correlated.