This post writes up tweets I sent last night.  The case is not overwhelming, and maybe not even decisive, but it’s worth rehearsing the arguments for.  I’ll write this as though it were for a UK audience only, but the basic idea applies, with some qualifications, for other central banks too.

In the UK QE meant the BoE creating new money, electronic central bank reserves, and buying long-dated government bonds [government bonds that mean that you get your money back in a long time, like 10 years] from private sector holders.

We can think of this as a two step process [even though it was only 1 step in practice].  Step 1:  create new electronic money and buy a short dated government bond [often called a ‘Treasury bill’].  Step 2:  find someone who will swap the short-dated bond with you for a long-dated one.

Step 1 is – a few details aside – what the central bank does when it does monetary policy in normal times, when the central bank interest rate is away from the zero bound.  This is the step that is pointless at the zero bound, and necessitates QE in the first place.  Unfortunately, it’s exactly the step that was emphasized by the BoE when QE was first done [see Mervyn King speeches from the time, and other BoE communication, including educational material about ‘injecting money’].  The fact that the interest rate is zero is telling us that people are sated with whatever it is that makes money special and distinct from government bonds, leaving them both to be default risk free assets with the same yield.  So the central bank undertaking that trade leaves things as they were.  And hence is not stimulative.  An attractive feature of emphasising step 1 was continuity.  ‘We used to do monetary policy using the price of money, now we are using the quantity, but it’s basically the same thing.’  Stressing continuity was a worthy motive, but not at the expense of misleading to the extent that you stress the bit of a policy you privately think does not work.

Step 2 is what might just do something/ did something in the past.  Long-term government bonds are thought to be less liquid than short term bonds.  They normally trade at a discount relative to what we would guess given expectations about interest rates over the life of the bond.  Buying them compresses this discount, bidding up the price, and lowering the cost of finance for private sector entities.

Since step 1 is redundant, QE could have been done with step 2 only.  This would, most practically, be carried out by the Treasury’s Debt Management Office, and involve them selling short term bonds and using the proceeds to buy long.  This is often termed a ‘twist’, after ‘Operation Twist’ in the 1961 carried out by the Fed.

Would there have been any advantage in doing a twist-only QE?

Obviously it would have cut out the misleading central bank verbage about pumping money. They surely would not have talked about pumping money if no money was pumped.

It would also have avoided fuelling the political difficulties created by the misapprehension of actual QE as being about funnelling money ‘to banks’, or other undeserving parties.

This was what provided the motivation behind ideas like ‘People’s QE’.  PQE involved the BoE creating money and using it to buy things ‘for the people’ and not ‘for banks’, like environmental infrastructure, or similar.  It was based on a complete misunderstanding of QE.  Somehow failing to realise that the money creation would simply put money in the accounts of other private sector entities [just like bond buying does, remembering that pension funds hold a lot of those], and which, obviously, end up being deposits at banks.  And it was a misunderstanding, or perversion of the money-creation part, which was always intended to be temporary, and to be reversed later.

This connects with the second advantage, which would have been to avoid antagonising the other constituency either wilfully misunderstanding QE, or using it to pursue objectives in bad faith:  the group Krugman refers to as the ‘permahawks’.  These saw the step 1 money creation as bound to lead to an inflation catastrophe.  And of course they were proven wrong – or have been thus far.

A related group that seems to include many otherwise reasonable people contends that although the money creation in step 1 is not in itself going to lead to inflation catastrophe, if only it were reversed as promised, in fact the promise to reverse is not credible.  QE is therefore helicopter money by stealth.  Although that kind of helicopter money is the least useful kind, since HM relies to some degree on how it influences expectations, this does not persuade these QE sceptics.  And their scepticism is fuelled by the perfectly reasonable accounts of how the post-crisis financial environment requires central banks not to shrink their balance sheets back to pre-crisis levels.  [See comments by Bernanke, Broadbent and others].  These look like excuses for wanting to pass off helicopter money as non-reversed QE.

If QE had been done as a Treasury twist in the first place, there would have been no money creation to promise to reverse, or partially reverse.  The necessary swelling of central bank balance sheets to accommodate a new post crisis demand for money would have happened separately and naturally.

A third advantage may have been to avoid a problem that Larry Summers worried about during the crisis, that at the same time as the central bank was buying up long term bonds, the Treasury was tilting issuance into long term, taking advantage of a cheaper cost of finance, but essentially undoing some part of the stimulus intended.  If the Treasury had been the agent tasked with implementing the twist, [selling short and buying long], it would have been forced to make sure – indeed the policy request would have been defined such – that the twist was over and above normal issuance patterns.

Although it would have solved some problems, implementing QE solely via a Treasury twist might have created others.

For starters, it would raise questions about whether monetary policy stabilization policy was properly delegated to the central bank.  Requests made by the central bank of the Treasury to undertake a twist can be turned down, or not implemented, or not implemented in full, or could be bargained over in private.  Or even if they are not, there might be the expectation that these problems would obtain.

In the UK this danger would be mitigated by the fact that debt management is undertaken by an entity that is somewhat separate from the Treasury, constituted as the ‘Debt Management Office’.

And it is also worth noting that how QE was actually done bound the Bank of England and the Treasury together anyway.  The Bank insisted on an indemnity for the financial risks entailed in QE, and so to all intents and purposes QE was done on the Treasury balance sheet anyway, not the BoE’s.  And it was operated under a sequence of letters specifying what could be bought, and up to what limits, written at the discretion of the Chancellor.  QE anyway relied on Treasury cooperation, cooperation that could, in theory, have not been forthcoming, or been withdrawn at any time, or only sustained in return for something.

If we put this in the context of broad recommendations that I have made, along with Simon Wren Lewis and Jonathan Portes, that at the zero bound the BoE should be able to advise/instruct the Treasury to design a fiscal stimulus [with conventional bond-financed tax cuts or spending increases] to replace monetary stimulus that it judges is missing, then having an agency of the Treasury do twist-only QE on its behalf looks more natural.  One has the BoE advising on the increment to the stock of net debt, and the other has it advising on the maturity structure.

An interesting thought experiment is:  Could ‘twist-only’ QE be done by central banks on their own?  The assets held by central banks tend to be off too short duration to begin with (almost necessarily, as they don’t want to be twisting back and forth as conventional monetary policy is fine tuned).  So the answer is:  Not without granting central banks the power to issue their own short term debt.  This is usually regarded as a non-starter, invoking the slippery slope argument ‘where would that end, what would they do with the money and wouldn’t that interfere with Treasury policy?’.  However, under adequately proscribed circumstances it surely could be done this way.  Undertaking a crisis operation using a new financial instrument would not have been ideal.   But even this objection could be overcome by having monetary policy done like this at all times.

This thought experiment gets us to one of the interesting questions surrounding QE, and in fact money and central banks in general:  if the central bank did issue its own short term debt, how differently would that stuff be viewed from government short term debt, and from money itself?

Answering those questions properly I am going to declare beyond the scope of this post, if not the entire blog, even if the presumption that QE, indeed conventional monetary policy, however operated, requires a provisional set of answers.

More prosaically, a few other points are worth making regarding how QE was actually implemented in other countries.

First, in late 2011/12, the Fed did do a twist-only round of QE.

Second, with regard to QE in the Euro Area, we have to remember that there would be 19 Treasuries to deal with.  The twist-only QE idea is not going to fly in a zone that prefers maximal central bank independence.  Credibility problems that exist between, say, the UK BoE and the Treasury, where the Treasury is headed by ministers who have recently reconfirmed the BoE mandate, are going to be magnified many fold when you have multiple Treasuries, none of whom may support the current ECB mandate, and some of whom may find it awkward to shorten the maturity of their outstanding debt, not an action normally contemplated when belts are already tight.

Third, the Bank of Japan has been piling into stocks, and long since abandoned BoE style QE as its sole focus.  What has twist-only QE got to do with them?

Three points.

The regular QE they did could have been done as twist-only.

Second, the stock purchases could have been made using bond issues rather than by creating money.

Third, taking a step back, the problem of regaining monetary control seems to have been for some time on a different scale there.  Even the most conservative of central bank oriented economists are tempted to conclude that helicopter money is a necessary tool for Japan.  In which case finding institutional wheezes to avoid the temporary money creation leg of QE is something of an academic exercise.  One hopes that they escape the zero bound with sufficient confidence that the idea once again becomes relevant, but we are not there yet.