The PiVe Capital Swap Spread Trades' case study (in Fixed Income Securities from Pietro Veronesi) suggested that in pricing a swap, using a swap curve appears to be the most reasonable methodology.
However, he later said that we use data of LIBOR rates for maturities up to 6 months and then bootstrapping to find and derive the swap rate for Z(t,t+0.5).
Based on my understanding, the swap rate will be higher than the LIBOR due to counterparty risk. However, considering most swap are transacted through dealers, should we be using LIBOR instead of the swap rate? I am not very sure on the rationale for using the swap rate.