If the panel is treated like a pooled cross section and you don’t take measures to control for individual fixed effects, you run the risk of obtaining biased estimates of the relationship between X and Y. The lines labeled with a D identify the pooled (and biased) OLS estimates.