Does Government Debt Management Matter? High-Frequency Identification From U.S. Treasury Quarterly Refunding Announcements

• Economics

• Answering Challenges to Advanced Economies

This paper measures the impact of surprises in the aggregate and relative supply of U.S. Treasury securities at different maturities on Treasury yields using a new measure of supply surprises. Surprises are measured as the difference between the supply announced by the U.S. Treasury Department at Quarterly Refunding Announcements and a novel dataset of pre-announcement expectations of supply from primary dealers. The reaction of Treasury yieldsis measured in a tight intraday window around the announcement, to precisely identify the causal effect of supply surprises on yields. We find that the unexpected supply of one percent point of GDP of ten-year-equivalent face value of debt in the next 3 months is associated with 0.07% to 0.20% increase in the 10-year minus 3-month term spread. This is larger than other estimates in the literature. The existence of supply effects is consistent with a simple theory of term premia adjusting to the total quantity of duration supply: imperfect risk sharing in an overlapping-generations macroeconomic model is sufficient to break Ricardian equivalence. Using maturity-specific issuance surprises, we show that supply shocks at short and intermediate maturities transmit strongly to longer yields, while long-end issuance has no independent effect on intermediate rates, a pattern that supports an additional role for preferred-habitat theories of segmented demand and imperfect arbitrage. We derive implications of these empirical effects in a version of a model of optimal debt management used by the U.S. Treasury Borrowing Advisory Committee. We find that recalibrating the magnitude of Treasury supply effects in line with our estimates lowers the optimal weighted-average maturity in favor of shorter-duration government debt issuance.