QE’s graceful exit

The ECB is expected to discuss this Thursday the future of its asset purchase programmes beyond the end of 2017. But exiting QE is no easy matter. The Fed’s policy normalisation plan has been three years in the making, with an update only last June, but the plan hasn’t been executed yet. The reason is the concern about the market impact of stopping or unwinding asset purchases. Just like bond yields have collapsed to negative values as a result of QE, they could rise uncontrollably when QE ends. And, in the case of the eurozone, the borrowing costs of peripheral member states might shoot up like they did at the height of the government debt crisis five years ago. I argue that a graceful exit from QE would be possible if central banks chose to act as market-makers, but they won’t for ideological reasons, and that the ECB actually can’t, for legal reasons.

When at the start of the financial crisis ten years ago Willem Buiter and Anne Sibert suggested that central banks act as market-makers of last resort, the central banks’ holdings of assets were negligible, and so in effect they could only act as buyers of last resort. Today, central banks are the single largest holders of government bonds, and so they could conceivably make markets by both buying and selling.

The implications of this for the QE exit strategy are profound. Instead of setting themselves a target for asset sales, central banks could simply sell their QE assets into market demand, such as it is at any given time. Insofar as the private sector has a preference for holding government bonds rather than cash, the central bank can oblige by selling some of its bond holdings and retiring some reserves, with negligible impact on yields. And, if liquidity preference returns and private investors would rather hold cash than bonds, the central bank can buy the bonds off of them in exchange for reserves, also stabilising yields.

In fact, that bond yields have collapsed as a result of QE, ultimately forcing the ECB to allow itself to buy bonds at yields below the (negative) deposit rate, indicates that the ECB’s asset purchases added reserves to the banking system when the extreme liquidity preference of the crisis years had alredy subsided. In other words, the ECB did QE too late,

But actually existing asset purchase programmes collapse yields on the way in and scare everyone on the way out precisely because they are quantitative. Central banks set themselves balance sheet targets, and execute them by buying a set amount of securities. And when you set yourself an amount and time to purchase, you have no control over the price, or indeed over the yield in the case of bonds.

It is likely that one major reason why central banks don’t act as market-makers is that it implies controlling the yield curve at all points, not just the short rate, and that goes against everything their economists believe. In the case of the Fed, yield curve control was abandoned as a result of the 1951 Treasury-Fed accord.

As to the ECB, it has had to defend its secondary market purchases of government bonds from the zeal of German critics who went all the way to the ECJ to stop them, and the ultimate defence was that they are not monetary financing of government debt because they allow the price discovery mechanism of the market to run its course. At least that was the argument by which the ECJ advocate general Cruz Villalón allowed Mario Draghi’s OMT policy to stand. But, in so doing, the European institutions have elevated market-led price formation to a governing principle of monetary policy. Yield curve control is out. And so is any hope of controlling the response of eurozone government bond markets to a steady unwinding of the ECB’s QE.