The residual sum of squares (RSS) measures the variance of residual terms in linear regressions. Now consider the two types of models we covered in class, e.g. the unrestricted model and restricted model: for instance:

y_t = b₁ + b₂x_2t + b₃x_3t + b₄x_4t+ u_t

y_t = b₁ + b₂x_2t + u_t

the first one is the unrestricted model and the second one is the restricted model under the restriction of ₃ = ₄ = 0. Without any calculation, which do you expect to be bigger: the RSS of unrestricted model or the RSS of restricted model? Why?