It's strange... when two parties enter into a swap contract, the initial agreement is worth zero, and the fixed rate payment recieved by the party long the swap is determined based on implied forward rates that would make the value of the floating contract equal to that of the fixed to create an NPV zero exchange.
However, implied forward rate movements almost always exceed that of the actual future spot rate as a function of the averaging effect of the increase or decrease. Does it make sense that swap contracts should be zero at the initial signing? If so, doesn't that mean that the party that long the swap contract almost always receive a fixed rate that is slightly higher than what should be received (provided that the six-month coupon implied by the forward rate overshoots that of the actual coupon paid in an increasing rate environment, and the long swap receives fixed payments higher than what is deserved) or receive a fixed rate that is slightly lower than what should be received (provided that the six-month coupon implied by the forward rate undershoots the actual coupon paid in an decreasing rate environment).
The floater that is implied in both scenarios derive its value from implied forward derived from zero rates. Unless the yield curve is a straight line, the value of the floater in the swap contract value equation seems almost always over/understated.