A few stream of consciousness points coming out of Thu Feb 12’s Inflation Report Press Conference.
The MPC are determined to ‘look through’ the effects of commodity price falls, and describe balancing risks either side of their effects. I don’t see that at all. There are presumably risks either side of their assumptions about the income effects of lower oil prices. But there is a decidedly downward tilt to distribution over outcomes for inflation expectations, and whether they fall, extrapolating the fall in headline inflation.
It was notable that we learned that the floor to interest rates has been lowered. Carney gave a very good answer to a question about this at the press conference. Previous worries that lowering rates too much would damage the balance sheets of banks and building societies had abated, because those balance sheets were now in better health. (I wondered too whether there was any change in the mismatch between the tracking of deposit and lending rates to Bank Rate, which was the driver for the projected balance sheet harm before). But when exactly did this analysis shift? The fiasco of the first forward guidance (was it a stimulus or not?) could have been avoided if the floor to rates could have been dropped back when Carney first took office.
It would have been nice to know what the floor now was. 0.25%, 0, <0? This could make quite a difference to longer interest rate forecasts, and the stimulus that could potentially got by allowing market forecasts of future rates to do the easing for MPC, should such a need arise. Given the new premium put on clarity and managing expectations of future rates, it is an ommission not to give more certainty about the new floor.
The revision to the floor suggests to me a determination not to resort to QE again. Which begs questions about MPC’s current views about the efficacy of the stimulus imparted by assets purchased thus far. For sure, MPCs view of its efficacy cannot have risen. The best guess is that the median member views this to be less effective. Is this manifest in the forecast?
Carney’s answers to questions on the impact of Grexit did not satisfy me. He described it as a hypothetical that can be reacted to as and when. Well, there are many hypotheticals built into the forecast. The entire edifice is a distribution over hypotheticals. The fact is, there are probabilities of things working themselves out in various ways regarding the negotiations over Greece, and these weigh on a rational forecast of EZ demand for UK exports and are manifest in all kinds of asset prices now. MPC can’t have coherent views about the other things in their forecast if they can’t disentangle the superimposing of Grexit risks on current and future data is doing to their outlook.
At some point, the dreaded ‘price level shocks’ term got a wheeling-out. This is a favourite of central-bank-economics-speak. But makes little economic sense. Let’s stare at a model like the one the BoE itself uses for forecasting. There are various kinds of shocks in that model. Shocks to technology, demand, government spending, perhaps financial spreads, tastes, with home and foreign varieties of all of these. Some of these shocks might affect relative prices, and some might not. Given the lags with which monetary policy works, and normal policy reactions, all of these shocks will affect the price level over shortish horizons. And what happens over longer horizons is up to the central bank. (In a macro model with flexible prices, a somewhat unrealistic but illuminating example, monetary policy works instantaneously, and what happens to the price level at any horizon is up to the central bank.)
So the label ‘price level shocks’ is highly misleading. It implies that there is just stuff out there moving this unrelated thing ‘the price level’ around that can’t be responded to, and luckily, since they don’t affect that other thing ‘the inflation rate’, don’t have to be responded to.
Whether shocks need responding to depends on what caused them, and whether, (as Carney pointed out, in fairness) they can be responded to quickly enough (before the effects of the shock dissipate, for example), and not on some bogus calculus about their being shocks that affect the price level or the inflation rate. In the future, I hope central bankers get booed if they use the term ‘price level shocks’.
At times like this, shortly before an election with great uncertainty about what fiscal plans are beyond it, great strain is placed on the – imperfectly applied – BoE convention of forecasting conditional on existing fiscal plans. This convention is there to avoid the embarrassment of having to forecast that the Government of the day would do something different from what it said it would. But it was often and rightly disregarded in some of its details. Because of course the lags between a policy change and its effects mean that the BoE need to know what will happen to fiscal instruments, not what is ‘planned’. So, how on earth is a distribution being formed over post-May-2015 fiscal variables? I’m intrigued. We’d need to forecast the election result (a coalition, but between whom) and how negotiations would tilt plans from those announced by each party.
Final gripe. The Report describes the current policy challenge as returning both inflation and capacity utilisation back to targets, objectives over which there is no trade-off. But we must surely think that financial sector impairment is still a significant factor, and, if so, that there IS a trade-off, ie that ideally we would be aiming for above target inflation to ‘over-stimulate’ the economy and make up for the financial constraints on the supply-side (which may explain some part of the very low productivity right now). (In general, the two-goal-variable characterisation of policy is far too simplistic. It works in the very simplest New Keynesian model. But in a model like the one the BoE uses, there are many factors to weigh up in figuring out optimal policy and it is not just a simple matter of weighing inflation and spare capacity. Spreads would enter, consumption and investment would enter separately, nominal wages, the real exchange rate….. the list goes on.)
longandvariable 