Ben Broadbent’s speech on the Bank of England’s Inflation Forecasts, and the yield-curve based interest rate trajectories on which they are based, is as lucid account of the issues as you will find, compulsory reading for interested observers.
Currently, the MPC present forecasts conditioned on constant interest rates [and constant asset purchases], and interest rate expectations derived from the yield curve. The custom has grown that one uses both to infer something about what will actually happen to interest rates. For example if inflation were forecast to glide pleasingly back to target under rising interest rates, but surge above it under constant rates, the custom is to infer that this means that the rising trajectory is preferred to the constant one.
Responding to the bouts of criticism that Mark Carney in particular has faced, Ben’s speech makes the following points.
First, genuine market expectations about interest rates cannot be cleanly extracted from market prices. [Perhaps this is particularly true now, with concerns about liquidity in short term sovereign securities.] So if we see that an inflation forecast conditioned under these ‘expectations’ seems to satisfy the mandate, or, as Lars Svensson used to put it ‘looks good’, that doesn’t mean that the BoE is going to follow ‘market expectations’. Markets may not expect what has been extracted by the Bank Staff from market prices.
This first point is a relatively minor quibble with the BoE-watching code of practice. At the point where we are trying to figure out what the BoE is going to do from their forecasts, no-one really cares whether the trajectories are based on cleanly extracted measures of market expectations or not. What matters is whether those numbers, contaminated with risk premia or not, look like interest rates that the BoE wants to set.
Ben’s second point – neatly exposited by simulations on the Bank’s forecasting model COMPASS – is that, measuring expectations aside, an inflation forecast trajectory that looks good can be achieved by many possible interest rate trajectories. So there is no way to tell, precisely, from an attractive inflation forecast, what will happen to interest rates.
This is true. Though there will, in fact, usually be ONE unique path that produces the best outcome for all the goal variables that the BoE cares about. That’s not just inflation, but competing measures of resource allocation, nominal wages, the real and nominal exchange rate, and more.
Ben’s comment does raise the issue, though, of the utility of publishing these conditional projections, if the paths on which they are conditioned are such bad indicators of what the Bank will do.
One purpose they serve, even if the interest rate projections were wildly out, is to illustrate the multipliers in the Bank’s model, as currently adjusted, massaged and baby-sat by its policy-making customers. If these multipliers were constant, there would be no need for constant reminders. But perhaps views change, and these charts allow us to keep track of that.
An even better way to do that would be to release and support working versions of the BoE’s model and its database, together with the forecast adjustments that allow one to scrutinise the forecast from quarter to quarter.
All in all, this speech reinforces to me the case for the Bank actually telling us what they think they would do to interest rates, given the way the economy is seen at the moment.
Models like the Bank’s forecasting model tell us that it’s not possible to make a coherent decision about today’s rate unless you have in mind how that fits into a sequence of future rates, a plan stretching far out into the future to the point where the economy is returned back to its state of rest.
Since these plans are not revealed to us, there are several ways to interpret MPC behaviour, and its communication strategy right now.
- MPC take a different view about how interest rates should be decided from the mass of work on monetary policy models, and DO ignore future rates when voting on current rates. This would be very concerning indeed. But perhaps they know something we don’t?
- MPC agree with this mass of work, and do formulate interest rate plans, but despite minuting many things in detail about their deliberations, they think it best to conceal these plans from us.
Going back to my time working with policymakers at the BoE, I can think of individuals who fell into both categories.
Those in category 1 tended to be those who did not have long-duration backgrounds in economics. That does not necessarily mean they must have been wrong. Who knows, our profession has been wrong about many things, so perhaps an exclusive focus on current rates can be justified.
Those in category 2 I think had four motivations.
One, not stated openly, so surmised by me only, was to avoid scrutiny and preserve maximum room for discretion and changing direction in a world where changes of direction would be interpreted by outsiders as incompetence. To the extent that is part of today’s mindset – I have no idea – we can obviously not expect the BoE to become more transparent of its own volition, and it would be for its Treasury owners, or Treasury Select Committee observers to intervene.
Another was the firmly held belief that these forecasts would not be understood as forecasts, and would be taken to be promises. Which, once broken, would, well, reveal the MPC to be promise-breakers. Who would be capable of anything, like disregarding the inflation target, perhaps. Or reveal them to be people who changed their plans, when the plan had not in fact changed, only the news about the economy.
A third motivation was the view that while the plans would be concealed, they could be promulgated in ways like those Ben today tells us they cannot.
A fourth motivation was that while, somehow, individually, MPC members could think about plans, it would all be ‘too difficult’ to arrive at one collectively. I always found this baffling. Committees in many walks of life have to agree multidimensional plans [plans for many things, like investment, prices, location, output, hiring….]. Surely MPC’s difficulties were not insurmountable, and no harder than the task of agreeing a collective sense of what the economy will do for a given interest rate trajectory – something they do already in the Inflation Report.
Today’s MPC clearly wrestles with these same issues and finds itself unresolved.
Despite Ben’s scepticism about what can be read into the conditional forecasts, they continue to be published, and probably partly because not all share the view that they don’t at all serve the purpose for which he says they are not fit.
Despite the concerns about interest rate forecasts being interpreted as promises, MPC members continue to talk verbally about such forecasts, through hints, with accompanying [correct] mini-lectures about the conceptual pitfalls, sometimes pitched as a telling off, if they feel their earlier hints were misinterpreted.
Perhaps it’s time for the Treasury Committee to intervene, and ask for some exchange of views about why, if these plans are formulated, we can’t be told about them. And, if they are not, why they are not.
The Bank might reasonably point out that the formulation and communication of its interest rate plans are an operational matter on which it should have independence. Grounds for compelling the BoE to change its practices would be the fact that the issue affects the ability of those who are charged with holding the BoE to account to do this job well.
At the very least, the BoE could be asked to subject itself to an independent review on this issue, permitting a frank exchange of views without any compulsion (other than the fear of shame!).