I have commented before on how bad the Bank of England’s fan charts are as a tool for communicating its forecast at the quarterly inflation report. The Bank half accepts these criticisms and has made some small but very welcome steps towards improving the presentation and communication of its forecasts.

Today, in the last post before I go on holiday, I want to spend some time looking at a more fundamental issue for next week’s inflation report: the forecasts themselves. We will have another go at this in the Financial Times next week and comments in advance are really welcome, but here are some very interesting preliminary results. Quick summary – the Bank’s forecasts are not very good. In fact they are shocking. And this matters because the Bank tells us that the forecasts form the basis for UK monetary policy.

Every August, the Bank does its own evaluation of its forecasting record and always pats itself on the back. This tiresome tradition arises since the Bank gives the forecasts extremely easy tests to pass. It compares its central forecasts with its own subjective range of uncertainty.  Here are some other very basic tests of forecast accuracy, which do not give such an encouraging result to the Bank’s vast teams of economists and forecasters.

Persistent optimistic bias

If you look at the chart, the Bank’s central forecast for growth  (mode market rate) is, on average, too pessimistic by 0.2 percentage points in the quarter it is forecasting and it gets more erroneously optimistic as the Monetary Policy Committee peers further into the distance. Read more