Why is there so much difference in a DID regression with region time specific trend between i.year and c.year?

[Linghui Han]

In a DID setting, I run the code "xtreg y i.eventdummy##i.treated x i.year##i.id, fe ice(cluster, id)" and got very small standard deviations for all the coefficients. However, I run the code "xtreg y i.eventdummy##i.treated x c.year##i.id, fe ice(cluster, id)" and got the normal results. Why are results so different?

[Linghui Han]

Maybe it is because too many dummy variables. The dataset is a panel with 30 regions in ten years.