Fudged Guidance, and the guess

Two more thoughts on the Fudged Guidance, which I think more accurately describes the launch of the new framework than 'Forward Guidance'.  Both are repeats.   The first rants about the 'Fudge', in the Guidance.  The second is about the unusual amount of guesswork that must have gone on to work out the effect on unemployment and inflation of a prolonged period of fixed interest rates.

The fudge, of course, is over whether the new state-contingent commitment to keep rates at 0.5 per cent should be considered a policy loosening or not.  In the Inflation Report press conference Mark Carney was careful to avoid saying in his prepared remarks that this is a loosening of policy.  Likewise, when pressed to clarify whether it was or was not in questions, he stressed that FG was about making the stimulus already applied 'more effective'.  The Inflation Report takes the same care;  no words on whether the overall stance of policy (considering all the instruments, and guidance on them) should be considered tighter or looser.  One is led naturally, though the text is careful not to say it explicitly, that the stance of policy is about the same as before.  Yet there is also text using those words in Carney's press conference and subsequent interview remarks 'more effective'.

Here is one interpretation [based on conversations with other BoE watchers].  Carney wanted looser policy, did not believe in delivering it via QE, but wanted to deliver it instead via forward guidance.  However, the majority on the MPC did not want looser policy, delivered through whatever means.  At least two MPC members (Ben Broadbent and Spencer Dale) had gone on record coming out sceptical of forward guidance.  The others one might presume must have thought about it and discarded this in favour of using QE instead.  However, rather than simply face the new Governor down, a compromise was struck.  Carney, who wanted looser policy, could find what he wanted in describing the existing stimulus as now being 'more effective' (ie more powerful, likely to stimulate demand, asset prices and jobs to a greater extent).  The other MPC members were happy with the new framework provided it omitted explicitly mentioning that policy was now looser, and presumably were happy with the 'more effective' language, since this could be interpreted not just as 'more powerful', but something like 'more skilfully done'.   The compromise must have been made more fragile by the rising temperature of data over the last two to three months.   The compromise is problematic for both, however.  Carney, who wants looser policy, would struggle to come up with an economically meaningful reason why he did not describe it as 'more stimulus', and must regret the obfuscation.  And the Hawks [for want of a better word]:  all the communication subsequent to the launch has emphasised that this is about trying to encourage more spending (and presumably attendant inflation) through underpinning confidence and reducing uncertainty.  How could an MPC member content with the previous level of monetary stimulus find the 'more effective' be happy with that?

In this sense this was Fudged Guidance.  A text that tiptoes around the key question of whether policy has loosened or not, allowing two presumably opposed factions to strike an agreement, but leaving both uncomfortable.

There is another interpretation.  We take all the communication at face value.  The majority of the MPC voted to stimulate the economy further by making the existing stimulus 'more effective'.  But, as I said in a previous post, why did those who were not voting for more stimulus (under the old rules where you could only do this with today's Bank Rate or QE) change their minds from the previous vote?  And if they did change their minds, why were they not content to make it clear that their guidance should be taken to be a loosening of policy?  Why risk being seen to have Fudged the issue?

If this was a compromise, it was unnecessary.  The previous Governor voted many times in the minority, and his ability to perform his role as chair of the MPC was not visibly affected by that.  The cost of compromising was to put at risk the effectiveness of the new communications policy.  It was also not terribly successful, (yet). Markets brought forward their expected date of a first rise in Bank Rate, when that first date was already before the MPC expect to raise rates.  So markets didn't see it as a loosening, or at least, if they did, they considered it one MPC would think better of later.  But the economics press did, and presumed that Carney had bent the rest of the MPC to his will.

Onto the forecast and the commitment itself.  What MPC did not tell you, either in Carney's verbal communications, or in the Inflation Report, is the following.  In trying to predict the effect of a prolonged period of fixed interest rates the MPC are in the realms of pure guesswork, and much more so than in normal times.  This is going to get nerdy, but it's important, because it is a key part of the overall picture in forming a judgement about how MPC have handled FG.  The model that the MPC uses to articulate their forecast and compute the effects of such policies is being stretched to its absolute limits.  And in two senses:  first, because the economy is (as the MPC see it) far away from its resting place, and second, because of the need to forecast under a protracted period of fixed interest rates.

Let's deal with these two problems in turn.  If the MPC is right that there is a large margin of spare capacity, the model is operating far away from its resting place.  But the further away from its resting place, the less accurate a model it is.  [Even presuming its the right model in the first place, which is another story].  A key feature of macro models like the one the Bank uses is that it presumes that prices are sticky.  If they were not, the MPC would have no real job, and certainly no influence over real things like activity and unemployment.  Dumping the right amount of money on the economy to achieve a constant k per cent would suffice.  A simplifying assumption made is that there is a fixed chance that firms change prices each period.  But so far away from the model's resting point, such an assumption becomes highly dubious.  Research by Peter Karadi at the ECB gives an example of how unreliable this assumption can be.  Hit by a big enough depressing shock, the economy starts to become very like an economy with flexible prices, as all firms find it worthwhile changing prices.  The appropriate policy in  a model that veers from being a basically sticky price economy to being a flex price one is very different from the right policy in a model like the BoE uses.

Now onto the second problem, to do with projecting under fixed interest rates.  This is dealt with in various papers, (eg papers by Gali, Blake) but most vividly by  Carlstrom et al.  They highlight the weirdness of these sticky price models by looking at what happens as interest rates are kept fixed for more and more periods.  At some point, the inflation response to an extra period becomes explosive.  And then further out, it can even change sign.  This is not to say that the real world is anything like this, and we should worry about the 3 year promise to keep rates fixed causing an explosion or an oscillation in inflation.  [Though if you were a New Keynesian literalist this is what you would worry about].  But it is a strong hint that there is something deeply amiss with the model, and that it is an extremely unreliable tool for speaking to the appropriate policy response to a shock that takes the economy down to the zero bound for a long period.  Now, the MPC have always explained that they are careful not to take the steer from models literally, and apply their judgement based on their immersion in economics or the real world more generally.  But they have not explained to us that the tools they are using are revealed by today's circumstances to be much more fragile than we guessed before, and that they are going on judgement [which could be drawing in profiles overwriting the model, or carefully selecting the model] more than previously.  Although MPC communications don't mention this problem, one presumes that they must be aware of it.  One can guess that they chose not to mention it for fear of reducing confidence in the policy measure.  But this course of action is a gamble, and it rather goes against the overall grain of the new framework for FG, [aside from the Fudge], which is to emphasise transparency.