Omission of variable in regression model

[Probability 0.99]

Hi, I am currently solving a question about the omission of variables in a regression model.

The question says that R^2 decreases after two variables are omitted in the model; I understand this part. However I don't understand why the restricted model will not have the lower SSE of 27108.
R^2= ESS/TSS. Here, I assume TSS stays constant...
When R^2 decreases then surely ESS will now be a smaller value. Why is it that SSE of the restricted model will be bigger than SSE of the unrestricted model ?

Thanks a lot.