Why money targeting would not be a good way to implement forward guidance

As the ECB slides inexorably towards a liquidity trap, debate focuses on whether it might be prepared to engage in a forward guidance policy that implied a Woodford-style inflation target overshoot.  Simon Wren Lewis suggests that a more palatable way for the ECB to implement forward guidance for its hyper-inflation scared Bundesbank veto-wielders [my qualification of the proper noun Bundesbank, not his] would be for them to make promises about the path of money.  If I recall correctly, Michael Woodford argues that money quantities could serve as a way to monitor a commitment to lower future interest rates, and correspondingly higher inflation and output.  So Simon's proposal is not without precedent.

However, I don't think it would be such a good idea, because all the old reasons why money targets were dropped apply with even more force at or in the proximity of the zero bound.

These reasons centre on the uncertainty about the quantity of money consistent with the desired inflation rate - or velocity shocks.  In the past, velocity - here another word for money demand - varied so much that it was deemed undesirable, perhaps even infeasible to meet the previously announced target, on account of the consequences for the real economy, or even for inflation.  I'm not old enough to have worked under the UK monetary targets, but those I worked for were, and they never tired of explaining that painful era.  Even the Bundesbank themselves, who Simon hopes his proposal will appeal to, amassed some pretty impressive cumulative misses of their own money targets through the 1970s.  [I recall a number like 50%?]

At the zero bound, our monetary models tell us that almost any quantity of real balances (money divided by the price level) would be consistent with the prevailing interest rates.  That's why Woodford explains that monetary injections at the zero bound (in the form of quantitative easing) would themselves have no effect at all.  And why the Bank of England erred in its early communication about quantitative easing when it sought to stress that the transmission mechanism was all about injecting money, and no different from before.  So in the zero bound period, a commitment to a particular level of money will not be very informative.  And would be completely harmless for a central bank that didn't want to change anything, provided it was reversed as the zero bound period ended.  Then looking forward, although the quantity of real balances consistent with nominal rates becomes determinable again, practically speaking velocity will be extremely uncertain.  And we need to add to this that - not articulated in the plain vanilla New Keynesian account of money and the zero bound - there will be a recovery in financial intermediation that will mean that the velocity of all broader monetary aggregates will be increasing, and by ex ante unknown amounts.  So there would be no telling what the level of any money stock consistent with the desired inflation overshoot would be.

I dislike strategies like this for other reasons.  The main being that the tactic of deploying an intermediate target again is confusing, even for the supposedly well-informed financial journalist community.  For example, Michael Woodford's suggestion at Jackson Hole that optimal monetary policy might be approximated by a nominal GDP target prompted a tirade of material by commentators who were for targeting nominal GDP for its own sake, (ie not for the sake of targeting inflation), and by some who didn't seem to get the distinction at all between intermediate or final targets.  This kind of commentary unhinges consensus around what it is the central bank should be doing and erodes the legitimacy of the current mandate.  In the UK, the papers were full of stuff suggesting that the inflation target should be set aside, which HMT thankfully, and elegantly, cut through in its review of the Bank's mandate back in March 2013.