MPC member Ian McCafferty has raised the possibility that the Bank of England might start to sell some of the £435bn of gilts it bought under its quantitative easing programme. When asked whether QE should be reversed, he said it was “a question that needs a bit of asking”. This raises a tricky issue – that policy is asymmetric, in the sense that the Bank selling gilts would not simply have the opposite effects to its buying of gilts.
Let’s start with best estimates. Bank economists have estimated that QE did indeed reduce gilt yields and boost economic activity. Martin Daines, Michael Joyce and Matthew Tong estimate that the first £200bn of QE cut yields in the 15-20 year maturity range by around 1.2 percentage points. And Martin Weale and Tomasz Wieladek estimate that QE equivalent to 1 per cent of GDP (£20bn) added around 0.18 per cent to real GDP.
There are two main mechanisms whereby this happened.
One is portfolio rebalancing. Bank of England demand for gilts raised their prices and reduced their yields. Some investors used the money they got from selling gilts to buy corporate bonds and equities which reduced companies’ cost of capital and so encouraged them to invest.
The other is a signalling effect. QE signalled to everybody that the Bank intended for short-term interest rates to stay low for a long time. This reduced gilt yields, because these should be equal to the expected path of interest rates over their maturity. And it stimulated spending (relative to what would otherwise have happened) by giving companies and consumers confidence that borrowing costs would stay low.
Which brings me to the asymmetry. Under QE these two mechanisms worked in the same direction. But under reverse QE, they will work in opposite directions and so offset each other.
The portfolio balance effect means an extra supply of gilts will depress their prices and raise yields. And this will have knock-on effects onto corporate borrowing costs which will depress capital spending. That’s simply QE in reverse.
The signalling effect is different, though. If the Bank announces plans to sell gilts, traders will see this as an alternative to rate rises – because monetary policy would be tightened via higher gilt yields rather than higher short rates. They will therefore reduce their expectations for short-term rates. But this will tend to reduce gilt yields – which will tend to offset the portfolio balance channel. The net impact of reverse QE will therefore be small, because the contractionary effect of the portfolio balance channel will be mitigated by an expansionary effect of lower interest rate expectations.
The stronger is the signalling effect, the more true this will be. And we have a good reason to think the signalling effect is strong. It’s that later doses of QE had smaller effects upon yields and output than the earlier doses. This is consistent with those earlier doses having a strong signalling effect and the later doses a weaker one.
It’s possible, therefore, that reversing QE won’t do as much to raise gilt yields as QE did to cut them. For the same reasons, the effect upon output might also be smaller.
There are, though, broader and more important points here. Monetary policy is a strategic game between central banks and the private sector with an uncertain outcome: this is a point made in a different sense by Andrew Caplin and John Leahy in one of my favourite papers. And policy can be asymmetric: a move in one direction is not necessarily the same as the move in the opposite. Economic policy is not a simple hydraulic process whereby pulling levers has stable and predictable effects. It's more complicated than that.
