Did central bank balance sheets really need to get so big?

Richard Barwell is head of macro research at BNP Paribas Asset Management

A decade and a half of quantitative easing has led to a massive expansion in central bank balance sheets. However, it is not obvious that the balance sheet had to get so big in the first place, or stay so big, given the range of views within central banks about how QE works and the reasons why QE took place.

Some policymakers believe that QE works by disturbing the balance between demand and supply in the bond market. For those individuals, the accumulated stock of purchases is a crude proxy of the current stimulus from QE so the balance sheet expansion was appropriate.

However, many — and perhaps most — central bankers do not attach much significance to this quantities mechanism. They believe that asset purchases only really influence asset prices when markets are dysfunctional and illiquid. In normal market conditions these policymakers believe QE influences asset prices primarily via a signalling mechanism and it is then less obvious why the balance sheet had to get so big.

If a QE announcement can shift investor beliefs and asset prices then it must contain new information, over and above what is already in the public domain, given current communications and past actions. But are any bond purchases strictly necessary in these circumstances? If central banks hold valuable private information about the state of the economy or the future conduct of policy then they could publish that information in a more effective and controlled fashion at a fraction of the cost and without any expansion in the balance sheet.

Rather than hinting that rates may stay lower for longer via QE the central bank could publish a path for the policy rate that clarified and quantified how and why rates would stay lower for longer. Alternatively, if central banks don’t hold back relevant information then there is nothing left for QE to signal.

Some academic economists counter that the balance sheet is the signal. They argue that the accumulation of a large bond portfolio fundamentally changes the internal policy debate and the future path of policy. Central bankers now supposedly have an incentive to keep rates low for longer: to avoid losses on those bonds. However, most policymakers would probably protest that they ignore such considerations and investors may query whether keeping rates low for far too long is necessarily the best way to avoid capital losses on long-term bonds.

Even if it was necessary for the balance sheet to get so big, central bankers could have prioritised reversing that process — quantitative tightening — in the past, present and near future.

Policymakers who believe in the signalling mechanism presumably accept that the value of the signal gradually decays over time. Once the market has deciphered the message in a QE announcement it is unclear what purpose the associated bond purchases serve.

The issue is whether QT would also send an unintended and likely unwelcome signal to the market. It is not obvious that it would, if central banks had made clear from the start that QT would automatically follow QE. All purchases could have been automatically and gradually unwound once the market got the message. Instead, central banks adopted a full reinvestment policy, keeping the balance sheet big until at least the first hike. Once that norm was established, deviating from it would have sent a signal.

Decisions around QT are more complicated for those policymakers who believe in the quantities mechanism. For them, any reduction in the size of the bond portfolio would imply a de facto monetary tightening. But central banks could have chosen to tighten policy through QT — shrinking the balance sheet — before they raised the policy rate. Instead, they chose to tighten primarily through rate hikes.

The standard justification for the current approach is that the impact of changes in the balance sheet on the economy is more uncertain than the impact of rate hikes so it makes sense to tighten via the more reliable instrument. However, you can perhaps flip that argument on its head. If the impact of asset purchases is so uncertain then it might make sense to remove that major source of macro uncertainty first.

All this discussion presupposes that QE was entirely a monetary policy operation. It was not. Irrespective of how you classify the operations that took place in March 2020 — market maker of last resort or risk taker of last resort — central banks were buying assets to first and foremost achieve financial stability objectives.

Different objectives behind bond purchases suggest that different rules of the game could apply. The default monetary policy assumption of full reinvestment until at least the first hike did not have to be adopted. Policymakers could have argued that they would buy and hold bonds for as long as markets remained dysfunctional but would gradually sell them back into the market as soon as market conditions permitted. Instead, the full reinvestment policy was adopted and the financial stability purchases effectively morphed into monetary policy operations after the fact.

Setting monetary policy without 20:20 hindsight is a hard thing to do. But the way central bankers think about how QE works and the reasons why they do QE suggest they might have been able to achieve broadly similar objectives with a significantly different path for the balance sheet. That is worth thinking about before the next time central banks buy assets for monetary or financial stability purposes.