According to the section on the CIR model in Hull's well known tome, "the general shape of the term structure at time t is independent of r(t), but does depend on t". (Here r(t) is the risk-neutral short rate at time t.) However, it seems to me that the entire yield curve structure in CIR is determined once you have determined r(t), i.e. the shape of the term structure does depend on r(t) in CIR, but doesn't depend on t, which is the exact opposite of Hull's conclusion. This can be seen by looking at the expression for bond prices in CIR. So is Hull's statement wrong?