Unwinding £ Quantitative Easing: the BoE Should Sell Calls

Julian D. A. Wiseman

Abstract: Some possible means of unwinding quantitative easing could worsen market disruption. Selling gilts by selling short-dated call options would avoid this, and is easily done.

Publication history: only at www.jdawiseman.com/papers/finmkts/20110206_ending_qe.html. Usual disclaimer and copyright terms apply.

Contents: Introduction (BoE holdings); Solution; Conclusion; Also See.

Introduction

As part of the extraordinary monetary operations resulting from near-zero interest rates, the Bank of England has bought a large quantity of gilts (table on right), in total about £178bn nominal, currently worth about £190bn excluding accrued interest. At some stage, probably after the policy rate has been raised from its multi-century low of ½%, the gilts will need to be sold. How could this best be done?

The BoE’s desiderata should include the following.

• Not to be robbed—to sell at a price consistent with the prevailing market price.
• Not to be a cause of disruption to the gilt market.

But “disruption” is a broad term. It has been reported in the press that some hedge funds are shorting OATs, issued by France (rated AAA), in anticipation of a weakening of its creditworthiness. Could the same happen to the UK? Of course the Bank of England would insist that she is not expecting any weakening of the UK’s creditworthiness. But nonetheless, if the market should be pricing same—mistakenly of course—then the BoE should wish not to be worsening the market’s stress. So imagine that the BoE has a pre-announced calendar of gilt sales, and that about the time of a sale, gilts are weakening fast. Would the BoE wish to be selling into such a crash? Surely not. Would the BoE wish to be contributing to a possible sense of panic by cancelling an auction? Yikes! Both are bad.

Solution

There is a simple solution, which would work as follows.

• By maturity divide the gilts into four sets, shown by colours in the table on the right. The 2013 to 2015 gilts are not to be sold: they are to mature in the hands of the BoE. The longer gilts are divided such that there are approximately equal amounts in each set: 2016 to 2020, 2021 to 2028, and 2030 to 2060.

• Once a week the BoE is to auction 28-day call options on one gilt from each of the three ≥2016 sets. The strike price would be at-the-forward, or perhaps, for simplicity, at-the-money. Typically the amount sold of each option would be £400mn.

• But which of the gilts in a set is to be the underlying? Within each set there is to be a rotation, each gilt taking turn. The ordering within each set should be somewhat jumbled, such that the shortest is not followed immediately by the second-shortest. The BoE should also allow the possibility, after some of the gilts are sold out, of later reorderings.

• If a set’s current gilt is nearly completely sold, then the quantity of options to be sold would be slightly adjusted. If the remaining quantity, less options outstanding, is >1½× and ≤2½× the standard option quantity (so £600mn to £1bn), options are sold on half of this amount. If it is ≤1½×, options on the whole are sold.

• Of course, for each gilt the quantity of options to be sold is capped at the remaining inventory, less the nominal quantity of options outstanding. If that means that no options on that gilt can be sold, the next gilt’s turn is advanced in the obvious way. (This can have no effect until there are only a few gilts in a set, the current gilt being nearly all sold.)

This has the natural advantages of selling options. If the price is falling, the BoE would not be selling, in the sense that the options would not be exercised. Further, the dealers’ delta-hedging of these options would make them buyers as prices fall, and sellers as prices rise, a gentle stabilising effect. (The stabilising would be ‘gentle’ because there would typically be only £4.8bn of options outstanding.) Both of the these advantages happen automatically, under their own steam, without any need for further decision by the BoE.

Given the numbers above, each week’s operation would sell, from each set, an average of 50% × £400mn = £200mn. As of February 2011 the BoE owns about £44bn of each set, so selling the whole inventory would take something like 220 weeks, or about four years, plus or minus some months. If the Monetary Policy Committee should wish to act at a different pace, the “£400mn” could be changed, before or after starting.

Conclusion

Some possible means of unwinding quantitative easing could worsen market disruption. Selling gilts by selling short-dated call options would avoid this, and is easily done.

Also See

A similar suggestion—sell gilts by selling calls—has been made to the Debt Management Office: see Methods for Distributing Gilts (December 2008), a reply to a DMO consultation; and a presentation of the reply given at the DMO in February 2009. These discuss the legal form that options should take (listed mark-to-market derivatives), and the same arguments apply to a program of sales by the Bank of England.