2. QE and QT: the balance sheet as a second lever
Macro 2: Central banks
Once the policy rate reaches zero, the traditional instrument runs out: you cannot cut much further the price of something that is already free. From that point on, central banks stopped operating on the price of money and began operating on the quantity and composition of the financial assets in circulation. That second lever is the balance sheet.
What it buys and what it pays with
Quantitative easing (QE) means the central bank buys financial assets in the secondary market, typically medium and long-dated government debt, and sometimes mortgage-backed securities or corporate debt as well. It pays by creating reserves: entries in the accounts banks hold with it.
Two misunderstandings are worth pulling out by the root.
- It is not buying debt from the state. Purchases are made in the secondary market, from investors who already held those bonds. The distinction looks technical, but it is the border between monetary policy and monetary financing of the deficit.
- It is not printing money that reaches your pocket. Reserves circulate only between banks and the central bank. The money people spend is created by commercial banks when they lend. If those banks do not lend, the reserves sit still.
What QE does, in a sentence, is a swap: it takes a long, safe asset out of the market and hands over an overnight asset in exchange. Someone who held a ten-year bond now holds cash, and has to decide what to do with it.
The three channels through which it bites
- Signaling. Buying ten-year bonds is an expensive way of committing yourself: if the central bank is going to hold that portfolio on its balance sheet, it has an enormous incentive not to raise rates abruptly. The market reads it as an implicit promise of low rates for a long time.
- Duration extraction. Level 1 explained duration. A long bond forces someone to bear interest rate risk; if the central bank takes it onto its own balance sheet, there is less term risk to share out among private investors, and the term premium compresses. This channel acts directly on the long end, which is exactly where the policy rate cannot reach.
- Portfolio rebalancing. The investor whose bond has been bought does not sit on their hands: they look for a substitute and buy corporate debt, or equities, or property. That shove travels right along the risk spectrum. It is the channel through which QE ends up showing in the price of assets the central bank never touched.
Why QE did not cause the inflation everyone forecast, and then did
Between 2009 and 2019 it was repeatedly announced that all this would end in hyperinflation. It did not: American inflation spent much of that decade below the 2% target. The reason is the one in the previous section: reserves are not spending. Banks were repairing their balance sheets, the private sector was deleveraging and credit was not taking off. QE inflated the price of financial assets far more than the price of consumer goods.
In 2020-2021 the outcome was different, and the difference lies not in the central bank but alongside it. Enormous QE was combined with direct fiscal transfers to households and firms, in an economy shut down by the pandemic and with supply chains broken. The money did reach hands that spent it, against supply that could not respond. QE on its own pushes asset prices; QE alongside massive fiscal transfers pushes goods prices. It is a distinction that saves you a great many wrong forecasts.
QT is not QE in reverse
Quantitative tightening (QT) normally means ceasing to reinvest maturing bonds, so that the balance sheet shrinks by itself. It sounds symmetrical. It is not, for two reasons.
- QE is announced in a crisis and QT in calm conditions. The first is a deliberate, attention-grabbing act; the second is designed to be dull, predictable and in the background, an automatic process nobody has to watch.
- There is a minimum level of reserves below which the payment system starts to creak, and nobody knows exactly where it sits until they get there. It is a floor you only discover by hitting it.
That is what happened in September 2019. With the policy rate at around 2%, the overnight repo rate shot above 5%, and some trades printed higher still. The Fed had drained reserves beyond what the system could tolerate and had to inject emergency liquidity. No crisis, no failure: pure plumbing.
Real case: the Fed's balance sheet, from under a trillion to almost nine
Before 2008, the Federal Reserve's balance sheet was around nine hundred billion dollars and consisted, in essence, of notes in circulation backed by Treasury bills. Nobody paid it any attention.
After the three rounds of purchases that followed the financial crisis, by 2014-2015 it was above 4.5 trillion dollars: more than a fourfold increase. A first attempt to shrink it between 2017 and 2019 brought it below 4 trillion, and that is when the repo scare arrived.
March 2020 wiped out that reduction in a matter of weeks. Over the following months the balance sheet went from about 4 trillion to more than 7, and peaked close to 9 trillion in the spring of 2022. Then came the largest QT in history, which brought the balance sheet back toward 7 trillion over the following two years.
Hold on to the asymmetry: going up took months and coming down takes years. The balance sheet stopped being an emergency tool and became a permanent instrument, and its size is now a macro variable in its own right, just like the policy rate.