A linear regression model is trained on panel data with repeated measurements from the same subjects over time. Standard OLS is used without accounting for the panel structure. Which two assumptions are most likely violated simultaneously?