8 May 2012
View speech in PDF format
The return of the surprise in central banking
In the old days, central banking used to be about surprises. When I studied monetary policy first, the distinctions were all between anticipated and unanticipated events, only the unanticipated mattered if policy was to improve on a supposed unemployment-inflation trade-off, so surprises were key.
Then along came Friedman and Phelps and gradually their insight that “You can’t fool enough of the people enough of the time” led to an approach to monetary policy that emphasised limited ambitions, predictability and – by and large – no surprises. I’ll come back to that.
But some of the great pivotal moments in the history of central banking are about shock tactics, the timing, their potential impact and the communications around them. These, I submit, are the three key lessons we need to keep in mind.
Timing is everything
When it comes to such surprises, timing can be everything – better not be late. Take this from an issue of the New York Times during the 1907 crisis: “Knickerbocker will be aided; President Barney quits and financiers promise to support the Trust Company; JP Morgan & Co. help”. But JP Morgan didn’t help enough and the Knickerbocker Trust Company suspended payment for 5 months. Next day the Trust Company of America was duly saved, but it was too late: “The fallout from permitting the crisis to surface spawned an avalanche of bank suspensions and a sharp contraction in the economy” (William Silber, 2007, When Washington Shut Down Wall Street). Notice the publication date 2007. I guess Silber was not all that surprised at the September 2008 sequence of events – Lehmans are let go, and, although AIG is rescued almost immediately thereafter, there is an ensuing panic- and credit crunch-induced deep global recession.
We may take this as the first lesson of central banking announcements: there may be no second chances: missing the moment for the decisive surprise intervention may place you in a very bad position.
Another great epoch-changing statement was that of Richard Nixon some 40 years ago, on August 15, 1971, ending the Bretton Woods regime of fixed but adjustable exchange rates. In that statement he remarked: “Let me lay to rest the bugaboo of what is called devaluation.” A bugaboo, according to Merriam-Webster is “an imaginary object of fear” or “something that causes fear or distress out of proportion to its importance”. Well imaginary or not, the ending of the Bretton Woods system did have consequences. (Though the consequences did not include the one feared by my future father-in-law who had just heard a few days earlier of the engagement of his daughter to a boy with a job offer from the IMF: would it survive?). Consequences there were: the closing of the gold window (announced by Nixon as temporary) became permanent and was followed by a decade in which prices in the US doubled. And they almost doubled again in the following decade!
So here is my lesson no. 2: surprise announcements can have long-run effects.
Credibility and truthfulness
Of course statements around devaluations and the run-up to them have had a tendency to weaken the credibility of central bankers (and other public officials). Take this quote from the Irish Finance Minister reported in the Irish Times, Thursday January 28, 1993: “I want to assure the people in London who are putting around the rumours that we are going to devalue today or tomorrow that they are wrong.” True enough, no devaluation occurred on the Thursday or Friday … but it did on the Saturday.
Although widely cited, former Federal Reserve Board Vice Chairman Alan Blinder famously did not say “the last duty of the central banker is to tell the public the truth.” What he did do, it seems, was conduct a survey of central banker attitudes, “in which officials had ranked ‘a duty to be open and truthful with the public’ last among criteria on why credibility was important.” Well, that’s even worse, though it’s not Blinder’s view; it was the view of many central bankers he surveyed in 1999. Devaluation denials may have been an obvious reason for this, and it is perhaps unsurprising that regimes such as the disastrous ERM, which create such pressures for disinformation, can be among the most ineffective. Lesson no. 3: central-banking-by-surprise makes candid communication extremely difficult. Yet we all agree that credible communication is essential for lasting policy success: so this is a problem with fixed but adjustable exchange-rate regimes, for example.
Floating exchange-rate regimes with inflation targeting (or some variant of it) are quite different in this respect – and I think that this is one important reason why their adoption has become relatively widespread. With inflation targeting (any variant, soft or hard), the sudden surprise announcement is not a prominent part of the policy arsenal. Instead the market’s expectation of official action (on policy interest rates, for example) is gently nudged towards the policymakers’ intentions by small steps, speeches, hints and graduated open market operations. A very different policy environment where, instead of dissembling before major actions, the whole aim of policy communication is towards having no surprises at all. Perhaps the most extreme example of this is the Federal Reserve pre-commitments of recent months.
Let’s think of the two major recent ECB policy surprises: the SMP in May 2010 and the 3-year LTROs in December 2011. How do we rate them considering the lessons that have been proposed above. The SMP announcement occurred at a time of great market tension. Sovereign risk contagion from Greece had just begun, and there was the additional background of the unnerving flash crash. The danger that sovereign spreads would spiral away, undermining monetary transmission as well as a lot of other things, seemed real, especially as the funding firewall that was quickly agreed at political level would take several months to become a reality. ECB intervention in secondary debt markets looked like a sensible and urgent action needed to nip things in the bud. A moment analogous to that seized by JP Morgan in 1907, if you will. For some, the SMP may have looked like something close to the dreaded monetary financing, but the Treaty is pretty clear on these matters, and buying plainly in the secondary market is not monetary financing. (Any student of the history of central banking knows how often central banks have been loaded up with unremunerated paper by spendthrift governments.
If central bank dealings with government paper might be portrayed as lying along a spectrum, the SMP lies well over to one end.) The action was prompt – the JP Morgan moment was not missed – and the spike of yields was removed. (cf. lesson 1) And what about the longer run? Well there have been consequences, but not those imagined at the outset. Sovereign spreads widened again and, though the program was used again with varying degrees of vigour, it cannot be said by any means to have eliminated the cross-country variations which are indeed impeding the smooth functioning of the monetary transmission. And it has not avoided the large debt exchange at steep haircuts for Greece, the original biggest target of the programme. Lesson 3 is I think confirmed. Communication around the SMP introduction, and the well-known differences of opinion which surrounded it, tended somewhat to undermine the coherence of communication by central bankers generally and was, I would say, problematic. The jury is still out on whether the policy has been a lasting success, and there are of course different views on what it could have been expected to achieve long-term.
Let’s hope that the 3-year LTRO, brought in once again in response to financial market pressures, not only on sovereign bonds, but especially on bank funding, does have the lasting effects that are no widely hoped for it. Certainly it was well-timed (lesson 1), though some observers will have hoped for this or more at an earlier stage, and can note that, despite its scale and duration, it has not brought, for example, Irish sovereign spreads below the levels that triggered the bail-out in November 2010. Let’s hope that lesson 2 applies, and that it will be of lasting value. As for lesson 3, this time the measure seems to have been relatively trouble-free on the communications front. But perhaps again it is too early to say, and the voices of the critics of large TARGET balances may eventually start to drown out more welcoming voices and ask awkward and destabilising questions about intra-zone credit risk. Instead, I prefer to think that the reason the LTRO was trouble-free was that it went with the grain: this action corrected the market’s misapprehension that the ECB was indifferent or “didn’t get” the existential worries that were prevalent in the market. As much as the technical aspects themselves, this favourable signalling function of the LTRO announcement reassured and stabilised markets.
I have ranged too widely for these remarks to warrant any general conclusions. The crisis has brought back centre stage the role of dramatic announcements with far-reaching effects. It is important to get the timing right: too late and a lot of damage can be done; too soon and the side-effects may be worse than the hoped-for effect (if the source of the problem has been exaggerated or misdiagnosed). And the communications issues are more challenging than the finely-honed practice of monthly announcements by inflation targeting central banks.