The writer is an FT contributing editor and global chief economist at Kroll
Quantitative easing has developed a certain resemblance to the Eagles’ “Hotel California” — you can check out any time you like, but you can never leave. We should pay more attention to quantitative tightening, suggest former Reserve Bank of India governor Raghuram Rajan and others in a recent paper. Commercial banks change their behaviour when there are plentiful reserves, making QT far more volatile and difficult to pull off than expected.
Our grasp of how QE and QT really work remains tenuous. In announcing a bond-buying programme, a central bank signals to the markets it is committed to accommodative policy and that rates will be low for a long time. The entire yield curve drops as a result. In purchasing long-dated bonds, the central bank pushes their yield down and in theory incentivises investors to move into higher return securities (the so-called portfolio rebalancing channel).
However, QT isn’t just QE in reverse. When rates are at the zero lower bound, the signalling channel is strong. But announcements about the central bank’s balance sheet are less effective when the policy rate is well above zero.
In 2017, Janet Yellen, then Federal Reserve chair, promised QT would be more “like watching paint dry”. The reality has been somewhat different. Rajan argues this is because commercial banks change their behaviour when the central bank expands its balance sheet, but do not change it back again when the balance sheet shrinks.
The mechanics of QE are a bit wonky. When the central bank buys bonds from investors, the proceeds are deposited in a commercial bank account. The banks steer the money into demand deposits (which can be withdrawn at any time) because they pay less interest than time deposits. To balance out these liabilities, the Fed credits the banks with the same amount of reserves as assets.
The reserves give banks confidence they can weather any significant deposit withdrawals, and they are also used to extend credit lines that generate fees. This shortens the average maturity of bank assets, undermining the portfolio rebalancing channel and increasing bank vulnerability to liquidity shortages.
According to Rajan’s data, none of this unwinds when the central bank shrinks its balance sheet and reserves become less ample. Instead, banks substitute lost reserves with other assets that are eligible collateral in repo transactions, to remain confident of getting enough cash if they need it.
But if every bank tries to transform their assets into cash simultaneously, there will inevitably be a shortage, as happened in the US repo market in 2019. Banks also continue to extend credit lines even as liquidity wanes, to maintain client relationships.
That means banks make greater claims on the system’s liquidity during QT, which may continue until there is a market blow-up. Central banks can step in and buy bonds again to paper over these liquidity crises, as they did in 2019, at the start of the pandemic and in the recent liability-driven investment freeze in the UK. But that ratchets up banks’ demands for liquidity still further — and makes QT even harder to pull off down the line.
One way around this is to minimise the signalling channel of QE, as the Bank of England did last autumn when it announced it would buy gilts for a very limited period, after the fallout from the Liz Truss-Kwasi Kwarteng “mini” Budget. But that would only work in a small-scale market meltdown. Imagine the Fed announcing in March 2020 that it would buy bonds but only for a short time, reserves would not be plentiful forever and rates would rise soon. Investors would have continued their dash for cash.
Central banks could simply forget about QT. Unlike commercial banks they can take losses and run in the red. But there are good reasons why they should not have an ever-growing balance sheet. Investors would have an incentive to take more risk. Governments may lean on the central bank to buy more bonds to finance pet projects. Central bank independence would be severely at risk, undermining credibility. A forever-distorted yield curve would make price discovery impossible.
Better bank capitalisation could help reduce vulnerability in the face of greater liquidity needs. Bank regulators could prevent reserve hoarding by allowing banks to meet an average of liquidity requirements over time rather than daily targets. Standing repo facilities can be extended to non-banks with good collateral, as the Bank of England has recently done. Ultimately, however, the best way to get out of QE may be not to start it in the first place. You don’t have to check out if you’ve never checked in.