Liquidity Value of Government Debt

Bank-created money, shadow-bank money, and Treasury bonds all satisfy investors' demand for a liquid transaction medium and safe store of value. We measure the quantity of these forms of liquidity and their corresponding liquidity premia over a sample from 1934 to 2016. We empirically examine the links between these different assets, estimating the extent to which they are substitutes, and the amount of liquidity per-unit delivered by each asset. Treasury bonds and bank transaction deposits are imperfect substitutes, in contrast to the finding of perfect substitutes of Nagel (2016). Bank and shadow-bank non-transaction deposits are closer substitutes for Treasuries, but provide less liquidity per-unit of asset than Treasuries. The results are directly relevant to the monetary transmission mechanism running through shifts in asset supplies, such as quantitative easing policies. Our results on the imperfect substitutability of bank and shadow-bank money also inform analyses of the coexistence of the shadow-banking and regulated banking system. We construct a new broad monetary aggregate based on our estimates and show that it helps resolve the money-demand instability and missing-money puzzles of the monetary economics literature.

We document that U.S. Treasury convenience moved positively with inflation during the inflationary second half of the 20th century but not before WWII or after 2000. A macro-asset pricing model explains this shift through two channels. Inflationary supply shocks raise the opportunity cost of holding money and money-like assets, endogenously increasing convenience yields. In contrast, exogenous liquidity demand shocks elevate convenience but depress consumption and inflation. Model estimates show that spikes in liquidity demand after 2000 account for the weaker convenience–inflation link and the emergence of negative bond-stock betas, distinguishing liquidity from non-liquidity demand shocks.

We analyze the term structure of Treasury liquidity premium (LP). Through a model where illiquidity shocks are alleviated by holding Treasuries, we show that LP term structure is shaped by expectations of future market liquidity, liquidity term premium, and Treasury supply. As predicted, the LP term structure is downward-sloping in recessions but upward-sloping in booms, and forward LP predicts future LP and market liquidity. Furthermore, LP is quantitatively important for monetary policy pass-through: LP dampens the pass-through of interest rate policy yet strengthens the pass-through of quantitative easings. We also use LP to infer the term structure of Treasury safety premium.