Treasury Markets and Financial Intermediation

How resilient is the U.S. Treasury market to foreign selling, inflation, and quantitative tightening? We answer this question by embedding an empirically estimated sector-level Treasury demand system in a forward-looking equilibrium model with risk-averse arbitrageurs who price expected future demand. Equal-sized foreign dollar retrenchments from different countries generate heterogeneous yield responses because countries differ in their maturity composition, yield elasticities, and macroeconomic sensitivity. Inflation affects long-term yields through opposing portfolio reallocations: falling foreign official demand amplifies the response, while the Fed’s demand response offsets it. Retrenchment by foreign official holders therefore reduces Treasury yields’ sensitivity to subsequent inflation shocks. At longer maturities, the yield impact of QT depends more on expected state-contingent Fed support than on the immediate reduction in the Fed’s Treasury holdings. Treasury-market resilience therefore depends on who absorbs the shock, whether it persists, and how policy responds, not on the size of the shock alone.

We document a regime change in the Treasury market post-Global Financial Crisis (GFC): dealers switched from net short to net long Treasury bonds. We construct “net-long” and “net-short” curves that account for balance sheet and financing costs, and show that actual yields moved from the net short curve pre-GFC to the net long curve post-GFC. Our theory shows the regime shift caused negative swap spreads and co-movement among swap spreads, dealer positions, and covered-interest-parity violations. Furthermore, the effects of various monetary and regulatory policies are regime-dependent. We highlight Treasury supply as a plausible driver of this regime shift.

We show that understanding the Treasury market requires both estimating granular investor demand and structurally modeling arbitrageurs. Using a new dataset of sector-level U.S. Treasury holdings, we estimate demand functions that exhibit strong cross-maturity substitution. Embedding these estimates in an equilibrium model with risk-averse arbitrageurs yields two main findings. First, Treasury market elasticity is steeply downward sloping in maturity, with very high elasticity in the T-bill market; without structurally modeling arbitrageurs, a pure demand system implies implausibly low T-bill elasticity. Second, cross-maturity substitution implies that monetary tightening raises term premia; without it, as in baseline preferred habitat models, the prediction reverses.

We demonstrate the passthrough of Treasury supply to bank deposits through bank market power. We show that a larger Treasury supply crowds out deposits with disproportionate effects in more competitive deposit markets. A larger Treasury supply further curtails bank lending and affects bank funding structure. The explanatory power of Treasury supply is not driven by other shocks to deposit demand and supply. In comparison, monetary policy rate hikes have a larger impact on deposit funding in more concentrated markets, consistent with the deposits channel of monetary policy. Our empirical findings are rationalized in a model of imperfect deposit competition.