The 2.5% Inflation “Knockout” is an Epic Disaster
This is how the MPC describe the inflation “knockout” – this is one of the conditions under which the current stance of monetary policy would be considered to be too expansionary:
On the basis of these considerations, the MPC has chosen to set its knockout at 2.5% at the 18 to 24-month horizon. While there is a range of views as to whether the knockout horizon should extend out to two years, the MPC’s best collective judgement is that an 18 to 24-month horizon strikes an appropriate balance between not bringing inflation back to the target so quickly as to threaten the recovery, while demonstrating the MPC’s determination to bring inflation back to the target over the medium term. The knockout is framed in probabilistic terms, such that it would be breached if, on average, it is more likely than not that the Committee’s projection for inflation 18 to 24 months ahead is half a percentage point or more above the 2% target.
The Bank publish all the data we need to assess past policy against this “knockout”, so I’ve produced a crude graph.
For each quarter (one IR forecast), we get three data points, looking forward 18, 21 and 24 months. Each data point is a probability representing “% chance that inflation will be above 2.5% in this quarter”. If any point goes above 50% that triggers the “more likely than not” clause in the knockout and a (possible) tightening of policy. Those data points are on the red, green and blue lines in the chart. Against that I’ve plotted annual nominal GDP growth, yellow line, on the right hand scale.
This is a particularly confusing graph, for which I apologise. What it says is that the indicators for the inflation “knockouts” have spectacularly bad timing. They spike up in mid/late 2008 when nominal GDP was crashing downward. They spike up again in mid-2011 when nominal GDP growth has slowed from 5% to 3%. And they spike up again in early 2013, when NGDP growth is still very slow at 2-3%.
The “knockout” actually triggers exactly once, in the February 2013 forecast, when the probability of the CPI rate being above 2.5% on the 18 month horizon is 51%. In fact Mervyn King eased policy in that meeting. Good thing he didn’t have a “knockout”!
This analysis is really telling us something we already knew: the MPC doesn’t like setting policy which produces forecasts of above-target inflation near the 2 year horizon. You could argue that the MPC would not have tightened policy on this basis in mid-late 2008, because the “knockout” does not actually trigger (go above 50%). But the salient point is that the MPC need an indicator which tells them they should have been dramatically easing policy by the middle of 2008; instead they will remain tightly focussed on the inflation forecasts.