In his letter to the Chancellor, Sir Mervyn says (pdf) 'it is'possible that inflation could fall back more sharply given the existing margin of spare capacity in the economy.'

That word 'possible' is doing quite a bit of work. My chart shows why. It plots core inflation (excluding energy, unprocessed food and taxes) against the unemployment rate, lagged 12 months, since 2002. If spare capacity tended to push inflation down, you'd expect to see an orthodox downward-sloping Phillips curve. But you don't. Instead, it looks as if there are two vertical curves. There the 2002-07 one, which saw unemployment around 5.3 per cent, and there's a post-2008 one, with unemployment around 8 per cent.

The data seems more consistent with the Nairu increasing in the crisis, rather than with spare capacity depressing inflation.

After all, we've had high unemployment for two years now, and yet core inflation, at 2.2 per cent on this measure, is above its 10-year average.

What's going on here? I suggest three elements:

1. There's a mismatch between spare capacity and demand. In Arnold's terms, the patterns of sustainable specialization and trade have been jumbled up. A newly unemployed civil servant in London, for example, cannot easily become an engineer in Aberdeen. Unemployment can therefore exist alongside rising cost pressures in some sectors.

2. The financial crisis was a supply shock. To see what I mean, consider why spare capacity should bear down on inflation. One reason is that it makes markets more contestable. Spare capacity enables a firm to expand quickly and cheaply. It thus encourages firms to cut prices in order to win more orders. Alternatively, it compels firms to hold prices down in order not to be undercut by potential rivals. But what if firms can't get the finance to expand or - what is overlooked - are unwilling to borrow for fear that the credit line might be later withdrawn? In this case, spare capacity won't encourage expansion and thus won't encourage price cuts.

3. Price wars are cyclical. In a famous paper (pdf) written in 1986, Julio Rotemberg and Garth Saloner showed that price wars were more likely in booms than slumps. The time to cut prices is when there are lots of customers to be won - and this is not in a recession.

If all this is right, then we shouldn't expect core inflation to collapse. Yes, overall inflation will drop for mathematical reasons as last winter's rises in VAT, food and petrol prices drop out of the comparison. But perhaps inflation will stay above target*. There is a middle way between the gold bugs who somehow think QE will be hugely inflationary and the agg demanders who expect inflation to slump.

All of which raises a question. If spare capacity isn't very disinflation, why does the Bank seem to think otherwise?

Here's a thought. Believing that a weak economy leads to significantly lower inflation allows you to relax monetary policy substantially whilst claiming to stick to inflation targets. But in fact, what you are really doing is targeting real economic growth instead. Maybe the Bank has been a closet nominal GDP targeter all along. After all, this - more than inflation - justified its superloose policy in 2008-09 and its recent resumption of QE.

* A caveat here is that the spare capacity might exist within firms, because of labour hoarding, and so not show up as unemployment.