When the policy rate hits zero (the "zero lower bound") the conventional lever is exhausted. You cannot cut below zero by much before people hoard cash. Quantitative easing (QE) is what central banks reached for instead, first in Japan (2001), then massively in the US, UK, and euro area after 2008 and again in 2020. It is the most misunderstood tool in macro, and getting it right is a genuine analyst edge, because half of what is said about it, by pundits and by professionals, is wrong. The key that unlocks it is M7.01's distinction between base money and broad money; hold that distinction and QE becomes simple.
How does QE work?
What QE mechanically is. QE is the central bank creating new reserves (base money, its own liability) and using them to buy financial assets, overwhelmingly long-dated government bonds (and sometimes mortgage-backed securities or, in extremis, corporate bonds), in enormous quantity. It is an outright OMO purchase (see the interest rate corridor) scaled up by orders of magnitude and aimed at long-term securities rather than short-term liquidity management. The central bank's balance sheet expands on both sides: assets rise (the bonds it now owns), liabilities rise (the reserves it created to pay for them). The US Fed's balance sheet went from under $1 trillion in 2008 to roughly $9 trillion by 2022 (check on FRED, series WALCL) — that expansion is QE.
Advertisement
What QE is trying to do (the channels). QE works, to the extent it works, through three channels, none of which is "printing money to spend":
- Duration/portfolio-balance: by buying long-dated bonds, the central bank removes duration (interest-rate risk) from the market, pushing up bond prices and pushing down long-term yields. Since long yields are what actually discount mortgages, corporate bonds, and equities, lower long yields ease financial conditions even with the policy rate stuck at zero. This is the main intended channel.
- Signalling: committing to buy bonds for a long time is itself a form of forward guidance: it signals the policy rate will stay low, reinforcing the guidance channel.
- Portfolio rebalancing / risk appetite: by swapping safe bonds for reserves, QE pushes investors who sold bonds to buy other assets (credit, equities), lifting asset prices broadly: the "reach for yield." This is also QE's most criticized effect, because it flatters the assets the wealthy own (see central bank independence).
Does QE print money?
What QE does not do: the reserves-versus-money point (read this twice). The folk story is "QE prints money, which floods the economy and must cause inflation." This is mostly wrong, and here is precisely why. When the central bank buys a bond from a bank, it credits that bank with new reserves. But reserves are money that only banks can hold and only circulate between banks and the central bank. Reserves are not broad money; you and I cannot spend a reserve; a company cannot pay a supplier with a reserve. Creating reserves does not, by itself, put a single rupee into anyone's deposit account. When the central bank instead buys a bond from a non-bank (a pension fund), the fund's bank does get a new deposit (broad money rises by that amount) and new reserves, so QE does raise broad money somewhat, by the portion bought from non-banks. But here is the crucial point. It does so by swapping one asset (a bond) for another (a deposit) in the private sector's hands; it does not make the private sector richer, and it does not force anyone to spend. The seller had a bond; now they have a deposit of identical value. Whether that turns into spending and inflation depends on whether the lower yields QE produces actually induce new borrowing and investment, which is a behavioural question, not a mechanical one.
What does the evidence on QE show?
The evidence bears this out and is worth carrying as a settled-but-nuanced view (be evenhanded; this is contested terrain): the 2008–2019 era of massive QE in the US, UK, euro area, and Japan produced persistently below-target inflation, the opposite of the hyperinflation the folk story predicted, because the reserves piled up inside the banking system (banks held them at the central bank) and broad-money growth and lending stayed weak in a deleveraging world. Then 2021–22 did bring high inflation. But the cleaner reading is that this time QE was accompanied by massive direct fiscal transfers (governments deposited money straight into households' and firms' accounts, which is broad money, created by fiscal spending, landing where it gets spent) on top of pandemic supply shocks. The natural experiment is almost too neat: QE alone (2010s) → low inflation; QE plus helicopter-style fiscal money plus supply constraints (2021) → high inflation. The lesson for the analyst is precise: QE changes the composition and price of assets far more than it changes the quantity of spendable money; it is a financial-conditions tool, not a money-printing machine, and the thing that reliably creates inflationary broad money is new lending or new fiscal spending, not reserves parked at the central bank. (This also quietly vindicates M7.01's endogenous-money view over the mechanical multiplier.)
What is quantitative tightening (QT)?
QT, quantitative tightening: the reverse. When the central bank wants to unwind, it does QT: it stops reinvesting maturing bonds (passive "run-off") or actively sells them, shrinking its balance sheet and extinguishing reserves. This drains liquidity and puts mild upward pressure on long yields. It is QE in reverse, though central banks run it far more gently (in the background, "like watching paint dry," as one Fed chair hoped) because selling long bonds aggressively can disorder markets. The analyst's watch-item: QT quietly removes a buyer (and a bid) from the government-bond market at the same time governments are often issuing more debt — a supply/demand squeeze that can steepen the curve and raise term premia (M7.03/M7.04). The size and pace of QT is a real, if slow-moving, macro variable; track the central bank's balance sheet (FRED WALCL for the Fed; the RBI's balance sheet in its weekly statistical supplement) as a dashboard indicator (M7.07).