The Bank of England’s Monetary Policy Committee raised interest rates for the first time today since 2007.
So here is a quick post-mortem on the November 2017 Inflation Report.
I would have voted against a rate rise today, for a few reasons.
First, it’s possible to read into the Inflation Report [and some commentators have, and anticipated this argument] that a reason to begin hiking is a change of view about the pace of growth of future output over the future. If this is the case, then this is not itself a motivation for tighter policy, unless for some reason the path of demand for a given policy is to be judged to have stayed where it was. But in the kind of models the BoE uses for forecasting, this would not generally be correct for a slow accumulation of potential output shortfalls. Demand would be expected to be correspondingly muted.
Second, the effect of Brexit on inflation via Sterling’s depreciation – expected as it is to be temporary, even if persistent – can be entirely discounted and the remit gives the MPC leeway to do that. [This was the justification – only in reverse – for tolerating the very large undershoot of the target that preceded it].
Third, if we strip out the temporary effects of the Sterling depreciation, we go back to the old position that the MPC are hoping to return inflation from below, to target ‘in a sustainable manner’. This was code language for discounting concerns that at the zero bound, when uncertainty is heightened, MPC should set policy so that the most likely outcome is to overshoot the target, and not to hit it. These arguments are old, but they were made lucidly by FOMC member Charles Evans.
Fourth, the pivotal judgement seems to be that slack is being used up, judging from extremely low levels of unemployment. [As MPC note in the Report, it fell to a 42 year low]. Yet this evidence is not as decisive as it first seems. Whole economy nominal earnings growth is flat at about 2%, and much lower at that than would be consistent with trend [1.5 plus 2 for inflation=3.5]. Private sector earnings growth has increased, but over a very short period, and from a very low base. The policy decision rests almost entirely on future growth materialising as a result of forces that take hold now and in the future. As the ever doveish Danny Blanchflower puts it crisply:
A few other points arising today and recapping after dripping them out onto Twitter hour by hour.
First, today reinforced my feeling that the MPC would be better to move to publishing unconditional inflation report forecasts that are based on what MPC feel is their preferred interest rate path. Today’s Press Conference seemed to have won over Chris Giles at the FT to this view.
This debate deserves its own post – and I was planning to write-up a slide show I did last month doing just this.
But, in brief….. the imprecision about what happens next – the intended pace and eventual resting point for rates – seems to confer very little or no benefit. And do no more than make it harder to figure out what they intend to do, and to judge subsequently whether MPC are being consistent in following through what they had previously intended or not. We must presume that the MPC have thought about when the next rate rise will come, and where rates will settle. And on what these views depend. Why not tell us?
One of the benefits was always said to be that MPC avoid presenting something that is misinterpreted as an unconditional promise [in this case to raise rates], which causes reputational damage and volatility when the non-promise is ‘broken’. But Carney and the MPC’s yield curve talking already incurs this cost. What else is stopping them?
Second, a nit-pick, but one that might be material. MPC condition their forecast on an average of eventual outcomes for our relationship with the EU, but on the particular assumption that we get to that outcome smoothly. Smooth transition is one possibility, but there is another class of possibilities – one of the variety of ‘no deal’s. So the forecast deliberately fails to weigh something that is distinctly possible [and in fact positively embraced by some prominent Brexiteers].
Missing this out of the distribution would be justifiable if either i) the range of implications for policy miraculously balanced out as neutral, or ii) it was ok to wait and see whether there was a deal or not before responding. i) seems highly unlikely, and anyway inconsistent with the rest of the IR forecast. ‘No deal’ is just a sharp and temporary version of the eventual outcomes, which are colouring the outlook for policy, as MPC are at pains to describe. ii) is a stretch: much better, at the zero bound, in the face of this uncertainty, to build in a response now.
But is this omission even internally consistent? Presumably the chance of ‘no deal’ is weighing on actual data as firms and households ponder the future. But MPC somehow have them slowly waking up to their assumed truth that transit will be smooth?
This odd treatment is an application of an old convention that manifests itself in other ways: the assumption that there would simply be no disorderly workout of the Eurozone crisis; the assumption that government will follow through with certainty on fiscal plans as they stand.
Forecasting under different assumptions is uncomfortable in each case, as it carries with it the risk that the MPC is impugning the competence or motives of other official actors. But not forecasting how everyone else sees it leaves the forecast necessarily disconnected from the policy decision [and indeed the market rate path used to communicate their intentions]. Not good!
A third point I wanted to make is about what the MPC did not do to tighten. They did not sell assets under the QE program. This is what they had said they would do all along. But I think it’s mistaken. QE-later was initially – and later, I think – justified as a way to avoid uncertainty caused by QE [an instrument whose effects we/they know less about than interest rates] being the marginal tool of monetary policy adjustment – eg in the event that economic conditions worsened again and there had to be a reversal, a subsequent repurchase of assets.
But this gets the uncertainty argument wrong. At least if my word-maths is correct. If the dominant stimulatory effect of QE is via its level – and such as the evidence is that it what we should believe – uncertainty is best minimised by getting back to the neutral level of QE [granted we are hazy about that] as quickly as possible.
Additionally, if you believe that QE has costs that are greater than costs of keeping rates away from base, then QE-second also neglects this argument.
The Econ press is naturally making a big deal of today’s decision, since rate rises are so novel, and some of the writers were probably at school when the last one was recorded. Opinion is inevitably divided [just as it is on the MPC itself]. I think the rate rise was premature, but it is hardly a mistake of historic proportions.
Monetary policy is still imparting almost maximum stimulus; it was always going to be withdrawn at some point, and gradually, and the question was just exactly when. If events confound the hoped for entrenchment of higher earnings growth, rates can be cut by the same amount that they were hiked today, and the reversal won’t have made much difference to the overall trajectory of the economy.
This icing-sugar time-series of Bank Rate compiled by the FT’s Gavin Jackson rather makes the point [that range of plausible alternative paths for Bank Rate is very small compared to its historical variation].