Treasury Debt Management
Since April 2024 the U.S. Treasury has retired $321 billion in outstanding securities through regular buyback operations. DurShift tracks every operation, computes the duration impact, and shows that par volume tells a very different story than duration redistribution.
Par retired + duration-weighted volume
The program ramped quickly from April 2024, with monthly par volumes reaching $30B+ by early 2025. The orange duration-$ line diverges sharply in months when long-dated liquidity-support operations are concentrated.
Share of total par vs share of total duration-dollars
The par/duration asymmetry is the central finding. The 0-2Y bucket retires over half of all par but barely a sixth of duration-dollars, while 10-20Y and 20Y+ together account for 23% of par but 63% of duration impact.
Cash management operations dominate par volume—$153 billion of short-dated securities retired almost entirely in the 0-2Y bucket. But the duration story is inverted: the 20Y+ bucket accounts for just 12% of par but 35% of duration-dollars, because each long bond retired removes 15-25 years of interest-rate exposure.
| Bucket | Par | % Par | % Dur-$ | Avg Duration |
|---|---|---|---|---|
| 0-2Y | $180B | 56% | 16% | 1.3 |
| 2-5Y | $45B | 14% | 12% | 3.4 |
| 5-10Y | $22B | 7% | 10% | 6.1 |
| 10-20Y | $37B | 11% | 28% | 13.8 |
| 20Y+ | $37B | 12% | 35% | 17.5 |
Around operations, front-end yields show modest declines (3M: -1.1bp, statistically significant; longer tenors: smaller and not clearly distinguishable from zero). Dealers increase net Treasury positions by ~$7 billion on average, with the largest accumulation in the 7-11Y sector.
These are simple averages, not causal estimates. Most buyback operations coincide with auction settlement dates, and overlapping event windows are common (70 of 121 operations have another operation within ±5 days). Confidence intervals are wide for tenors beyond 6 months.
Par breakdown by security type
Nominal coupons dominate at 97% of par. TIPS buybacks ($11B) are smaller but analytically distinct—they use real yield proxies from the H.15 inflation-indexed CMT series and carry higher duration per dollar of par.
Average change day before → day after · only 3M and 6M are statistically significant
Only the 3-month and 6-month tenors show statistically significant declines. The long end (20Y, 30Y) trends slightly positive. This pattern is consistent with cash management operations reducing short-dated supply while not mechanically suppressing long yields.
Average change in primary dealer Treasury coupon positions
Dealers increase holdings across most of the curve around buyback operations. The 3-6Y bucket is the only sector where positions decline on average, possibly reflecting where Treasury concentrates purchases in that maturity range.
Size = par accepted · hover for details
Two clusters emerge: liquidity support operations (blue) sit at high duration / low par, while cash management operations (orange) sit at low duration / high par. The separation reflects the program's dual mandate.
Top 10 CUSIPs by total par accepted
| CUSIP | Type | Coupon | Maturity | Total Par | Operations | Avg Price |
|---|
Duration-dollars per operation · blue = liquidity support · orange = cash management
Liquidity support operations (blue) produce the tallest bars because they target long-dated, high-duration securities. Cash management operations (orange) are frequent but carry low duration impact per operation.
SOMA portfolio composition · monthly from NY Fed
The Fed's SOMA portfolio peaked above $8T in 2022 during QE and has since declined through quantitative tightening. Treasury buybacks operate against this backdrop—when the Fed is also reducing duration supply, the net impact on publicly held duration is amplified.
For each buyback operation, DurShift downloads the official results XML from TreasuryDirect, extracting per-CUSIP accepted par, coupon rate, and maturity date. Yield proxies are interpolated from the Federal Reserve H.15 constant maturity term structure—nominal CMT for coupon securities, real CMT for TIPS. Modified duration and DV01 use standard semiannual cashflow discounting. Duration-dollars = market value (par × price/100) × modified duration.
Event windows compare dealer positions (NY Fed primary dealer statistics) and Treasury yields (H.15) in the days surrounding each operation. Nearby gross issuance is estimated from Treasury auctions within a ±14 day window (not matched replacement supply).
Every input is public and verifiable