Traditionally US swap spreads were traded as LIBOR or OIS swaps versus USTs.
In the former case the spread at the short end of the curve was very much a function of LIBOR repo spreads. Further, LIBOR which was a measure of bank credit was fundamentally different from the credit of the US government. Therefore, some value of these short terms spreads could also be ascribed to this difference in credit.
In todays world, most swaps are cleared and therefore the idea of these swaps defaulting is fairly close to 0. Further, with LIBOR now being redundant, we have SoFR swaps versus cash as the benchmark spread. Therefore, barring specials the index on the swap is a fair proxy for funding on the bond at GC.
If we assume SoFR is representative of GC rates, arguably at the short end of the curve, Matched maturity ASW should be close to 0? In fact the only thing which could be driving this away from 0 would be the balance sheet capacity of banks..which implies longer spreads should be negative (if measured as swap over bond yield).
Therefore, in an ideal world where bank balance sheets were not impacted by regulatory bounds and were hypothetically abundant, the new swap spreads should be close to 0. The fact that long end spreads are materially negative shows or implies this b/s scarcity ?
Are their any other subtle factors outside of balance sheet which determine these spreads today (ofc treasury issuance is just a corollary to this point, in that high levels of UST issuance would cheapen certain sectors as balance sheet is constrained).