My claim that expectations are a weak lever for policy has been questioned. Nick says:

Expectations matter a lot for the demand for long-lived assets, and hence for the price of long-lived assets. The lower the rate of interest, the more expectations matter. What happens this year doesn't matter very much at all (unless it affects expectations of future years), and it matters even less as the rate of interest gets lower.

This seems reasonable. But it has an implication which doesn't always fit the facts. It implies that as long-term real interest rates fall, the price of "growth" stocks should rise relative to value stocks. This is because growth stocks - by definition - are expected to offer more future cashflows than value ones, and lower interest rates raise the value of future cashflows. But my chart shows that this isn't always so. Between 2007 and 2010, growth stocks did indeed out-perform as index-linked yields fell. But since 2011 they have not done so. This tells us that the link between longer-lived assets and interest rates isn't as simple as theory tells us. This might be because the stock market is short-termist (pdf) and so longer-term expectations just don't matter much.

This, though, was not the main point I had in mind. I was thinking instead of my second chart, which shows NOP's survey of the public's expectations for inflation in the following 12 months.This chart shows three things:

1. Expectations are not well anchored to the 2% target. Instead, as Adam Posen has said (pdf), they are determined more by the recent data. This poses the question: if the inflation target does not much affect short-term inflation expectations, why should we suppose that an NGDP target would much affect short-term NGDP expectations?

2. Inflation expectations have been above 3% since May 2010. This implies that the public has expected sharply negative real interest rates. But this has not led to the rise in consumer spending that theory predicts.In January 2011, the consensus expected real consumer spending to grow 1.2%. In fact it fell 1.2%.

3.Nor have high inflation expectations led to higher wage growth; average earnings have risen 1.4% a year in nominal terms in the last two years.

These three facts imply that there are two weaknesses in the expectations lever; the Bank cannot easily affect short-term expectations, and such expectations don't obviously have the effects on macroeconomic aggregates that one might think.

Now, I agree that, in theory, policy credibility matters. And I don't doubt that there are times and places when policy-makers can affect expectations; I suspect they have more influence upon financial market expectations than real people's. I fear, though, that economists who invoke "expectations" and "credibility" are making the error of mistaking the tidy maps of models for the messy terrain of reality.

I do not say this to reject NGDP targets. Instead, it is to say that - insofar as these have merit for the UK economy here and now - it is because they would lead to a greater monetary stimulus, rather than because they would affect expectations and hence behaviour.