Quantitative easing (QE) is an unconventional monetary policy tool in which the central bank purchases large quantities of government bonds (and sometimes other assets) to push down long-term interest rates when the short-term policy rate has already reached the zero lower bound.
The Fed conducted three major QE programs after the 2008 financial crisis:
During COVID-19, the Fed launched another large program, purchasing 120 billion per month in Treasuries and agency MBS starting in March 2020. Combined, QE1 through QE3 added roughly 4 trillion to the Fed's balance sheet.
The core transmission mechanism is duration removal. When the Fed buys a 10Y Treasury from a pension fund, the pension fund receives cash. Cash carries zero duration. A pension fund with long-dated liabilities cannot hold cash indefinitely. It must reinvest in long-duration assets to maintain its liability match.
The pension fund then buys what is available: investment-grade corporate bonds, agency MBS, long-dated municipal debt. Demand for those assets rises, their prices rise, and their yields fall. The effect propagates across the credit spectrum without the Fed purchasing a single corporate bond.
This chain explains why QE compressed credit spreads and mortgage rates, not just Treasury yields. The Fed extracted duration from private portfolios. Private investors restocked duration from whatever remained.
The NY Fed's Adrian-Crump-Moench (ACM) model provides the cleanest read on QE's impact. Before the 2008 crisis, the 10Y term premium ran at approximately +150 basis points. By 2015 and 2016, after three rounds of QE, the ACM model placed term premium at roughly -100 basis points, a compression of approximately 250 basis points over seven years.
Research by Gagnon et al. (2011) and Li and Wei (2013) at the Federal Reserve estimates that QE1 through QE3 reduced the 10Y Treasury yield by approximately 100 to 150 basis points relative to the counterfactual path without asset purchases. The term premia tool on this site plots the ACM series in real time, so you can see how term premium has evolved through each policy cycle.
Conventional rate cuts work primarily through the front end of the curve. The Fed lowers the federal funds rate, short-term yields fall, and the whole curve tends to shift down as forward rates re-price. The transmission is fast and direct.
QE operates differently. It specifically targets long-end yields and term premium. When the Fed is already at zero, as it was from December 2008 to December 2015 and again from March 2020 to March 2022, the front end offers no further room. QE is most powerful precisely in that environment, because it reaches the only part of the curve still available for compression.
The practical implication for portfolio managers: QE flattens the curve by pulling the long end down while the short end stays pinned at zero. Curve steepeners funded in 2s10s become painful trades in an active QE environment. The yield curve tool lets you overlay historical curve shape against Fed balance sheet milestones.
A large QE balance sheet must eventually be unwound through quantitative tightening. The exit is technically and politically difficult. The taper tantrum of 2013 illustrated the market sensitivity: when Fed Chair Bernanke signaled in May 2013 that tapering was under consideration, the 10Y yield rose from approximately 1.63% to roughly 2.98% by September 2013, a move of approximately +135 basis points, driven almost entirely by a repricing of term premium rather than changes in rate expectations.
QE's transmission to the real economy also depends on conditions outside the Fed's control. If banks are unwilling to lend or businesses are unwilling to invest, compressing term premium does not automatically stimulate activity. The portfolio balance channel operates through financial markets first, and the connection to lending and output is indirect.
For the current unwinding dynamic, see quantitative tightening. For live term premium data, the /premia tool tracks the ACM model daily.
Does QE cause inflation?
QE expands bank reserves, not necessarily broad money supply. Whether it causes inflation depends on whether banks lend those reserves and whether the economy has slack to absorb additional demand. During QE1 through QE3, core PCE inflation remained below the Fed's 2% target throughout the purchase programs. The COVID QE episode coincided with supply shocks and fiscal transfers that are harder to attribute to QE alone.
How does QE end?
The Fed can stop reinvesting maturing principal, allowing the balance sheet to shrink passively. It can also sell assets outright, though active sales are rare. The pace of balance sheet reduction matters for markets: faster runoff removes more duration per month and puts upward pressure on term premium. See quantitative tightening for the mechanics.