The Portfolio Balance Mechanism (PBM) theorizes that reducing the supply of U.S. Treasuries (USTs) increases their prices and prompts the creation of similar assets for preferred-habitat investors. The PBM is recognized as a possible mechanism to explain various important phenomena. We identify the PBM using the suspension of 30-year UST bond auctions starting in 2002. The suspension raised long-term UST prices and led to the issuance of safe, long-term collateralized mortgage obligations (CMOs) created by tranching mortgage pass-throughs to satisfy habitat-preference investor demand. The heterogeneity of UST and CMOs results in an unusually clean and unambiguous identification of the PBM.

As investors adjust portfolios at different speeds, quantitative tightening shifts the composition of the marginal investors, which makes the Treasury market fragile. We show empirical evidence for the shift in investor composition, and develop a general equilibrium model of bond valuation with heterogeneous investors to understand its implications. In the model, long-term bond investors have higher risk-taking capacity, but face a portfolio adjustment cost; liquidity traders have lower risk-taking capacity, but can trade freely. Our model predicts a novel overshooting pattern: when the central bank unwinds its bond purchases, slow adjustment by long-term investors requires liq- uidity traders to absorb the imbalances, who demand a high risk premium and lower the bond price below the long-run level. As a result, quantitative tightening is not simply a symmetric reversal of quantitative easing.

The paper develops a general equilibrium model in which bonds serve both as investment opportunities and as collateral for loans. The model integrates the yield curve with the repo market. Preferred-habitat investors and arbitrageurs generate downward-sloping demand for bonds and upward-sloping supply of collateral. The results show that quantitative easing increases the relative collateral value of targeted bonds in the repo market, (i) strengthening local supply and (ii) dampening term premium effects of asset purchases. Academics and policymakers should consider the bond and the repo markets in combination since traded assets and economic agents are the same in both markets.