Expertise: King's red herring

Why should the Bank of England have operational indepdendence? I fear that in his Adam Smith lecture (pdf) on Sunday, Mervyn King may have mis-stated the argument. He said:

Responsibility for setting interest rates has been delegated to a group of people - the Monetary Policy Committee - with the appropriate technical competence...Expert judgment is needed because changes in the way the world works mean that monetary policy cannot be run on auto-pilot.

I've got two problems with this.
1. In theory, the Chancellor could set interest rates by drawing on as much expert advice as the MPC has. The MPC doesn't, in principle, have an advantage in expertise. It has an advantage in credibility. Markets trust the MPC more than it trusts politicians to hold inflation down. Indeed, it's quite possible - as Paul Ormerod argues here (pdf) - that no feasible amount of expertise would allow us to predict and control inflation.
2. In practice, monetary policy does seem to have been run on auto-pilot for most of the time since 1997. The chart shows that interest rates have closely followed a Taylor rule (more here) since then*. The two main deviations were in 1997-98, when policy was tighter than the rule warranted, and in 2002-03, when it was looser, possibly because the Bank took out insurance against a protracted US downturn and US deflation - just as the Fed did. It's not obvious that it required great expertise to follow this path, or that a Chancellor would have done much different; coincidentally, the excess tightness came straight after the general election.
* The rule I used is (0.5 x output gap) + (1.5 x RPIX inflation) + 1.6. This means rates should be 5.4% when output is at its potential level and RPIX inflation is 2.5%. I'm using the Treasury's estimate of the output gap.

Comments

[Dave Petterson]

I think that although the politicians could follow the simple rules there would be undue attention paid to additional, political, issues that would occasionally screw the results.

In the 1980s I installed a forecasting module into our ordering system. It forecast our orders based on seasons and history. On the bench it kept our inventory down to a minimum with no shortages and when it went live our inventory only dropped a few %. After a few weeks with no real inventory decrease we examined the figures and discovered that the warehouse manager was taking the figures from the system and 'fixing' them prior to order because he was concerned about running out and didn't trust the computer. Once he was 'fixed' our inventory dropped so much we sold the warehousing site and kept a small warehouse on the manufacturing site.

Human intuition and experience impacts on the systems and some of us have hidden agendas which to those on the outside means it does not make sense.

[Mrs Trellis]

Could just be that your Taylor rule is close to the MPC benchmark - to operate it will have evolved a benchmark which balances output and inflation deviations, and probably based on a rather simplistic expectations formation model. A behavioual model of the power plays between MPC factions might be sufficient to explain the deviations - or it could just be that the data has been revised since then.