Mark Carney appeared at the Treasury Select Committee today for interrogation before being confirmed as the next Governor of the Bank of England. The big question everybody wanted answered is whether he favoured a move from inflation targeting to NGDP level targeting. The answer is ‘no’, but the reasons are interesting.
Carney is a known proponent of central bank commitment and the use of forward guidance. In recent speeches he has also spoken favourably of NGDPLT and that has prompted a storm of commentary in the UK; little of it was favourable towards the idea. Consequently, all eyes were on Carney’s evidence today. Reading the comments on Twitter suggests that he dismissed NGDPLT and retreated from his previous statements. I don’t think that is true at all.
In both his oral and written evidence he called flexible inflation targeting “the most effective monetary policy framework implemented thus far.” However, he was at pains to point out that there are problems with it at the ZLB and there are other potential regimes, such as NGDPLT, that might help in those circumstances. In his oral evidence he spoke at length about the benefits of commitment and history dependence when encountering the ZLB. Despite that, he was not in favour of an immediate move away from inflation targeting, as you might expect given the outcome of the Bank of Canada’s review. Some of the reasons he gave are well-known: the problem of revisions and data quality, for instance. Notably, he did not think that NGDPLT would unhook inflation expectations and commented on the additional credibility a central bank could gain by implementing an inflexible rule.
The most interesting argument he made against level targeting was the one he dwelt on in his oral evidence: it relies upon people having rule-consistent expectations. That is to say, the success of a central bank relies on people expecting that it will implement its stated plans, and behaving as if they will come to fruition. Of course, he did not make the naive argument that people’s expectations are irrational. Rather, he pointed out that expectations among the populace have inertia and take time to change. If a large portion of the population have persistently incorrect expectations following a change in target then it would be costly in terms of welfare. He alluded to agent-based modelling done by the Bank of Canada to claim that these transitional costs as expectations gradually adjust could outweigh the gains to the switch.
In summary, he thinks NGDPLT is a great idea but hard to put into practice (data issues) and costly to implement (transitional costs of changing expectations.) Relative to the commentary in the UK press that is a ringing endorsement: one of the top central bankers in world says that the only real barrier is the details of implementation.