News that three MPC members voted to expand QE at the February meeting surprised markets, leading some commentators to claim that the Committee’s “reaction function” has changed to allow higher inflation. While this is possible, the vote was consistent with the prediction of the “MPC-ometer” model followed here – it was, in other words, explicable by recent data.

The MPC-ometer predicts the outcome of each month’s meeting based on the latest values of 12 economic and financial inputs, including business survey activity measures, the quarterly GDP change, manufacturers’ price-raising plans, consumer inflation expectations, credit spreads and equity prices. The forecast is in the form of the “average interest rate vote” of the Committee’s members; for example, if four vote to cut Bank rate by 25 basis points (bp) while five prefer no change, the average vote is -11 bp (i.e. four-ninths of -25 bp). The February prediction was -9 bp, consistent with three members voting for a 25 bp rate cut. The same number, instead, sought to expand the QE programme by £25 billion.

The February forecast was influenced by the weak fourth-quarter GDP estimate and also took into account the tendency for policy changes to occur in Inflation Report months. The dovish impact on the model of the GDP number will drop out next month, although the bias revealed by the three February dissents works in the opposite direction. With seven inputs available, the current prediction for March is -4bp, suggesting that a majority will again vote for no change. The final reading will depend importantly on any fourth-quarter GDP revision and February consumer and purchasing managers' surveys. (The chart incorporates the current March estimate.)