A model of the MPC’s historical “reaction function” suggests that the Committee will cut Bank rate by 25 basis points (bp) to 0.25% at its August meeting, based on current economic and financial information. The model forecast would be consistent with several members (Haldane, Vlieghe?) voting for more significant action but others (Forbes, McCafferty?) favouring “patience” or judging growth / inflation risks to be balanced.

A further possibility is a smaller rate cut (12.5 bp?) coupled with a modest QE programme (£30 billion over three months?) – such an approach has merits given uncertainty about the benefits / costs of moving rates towards zero. A less likely but more radical option would be for the MPC to move to an ECB-style dual-rate system, paying a “Bank rate” of 0.5% or 0.25% on the bulk of banks’ reserves but imposing a zero or negative rate on the top tier.

The “MPC-ometer” model attempts to explain the “average vote” of Committee members using information known before the start of each month’s meeting. If five members vote for a 25 bp rate cut and the other four for no change, the average vote is -14 bp (five-ninths of 25 bp). QE is allowed for by assuming that a 25 bp rate cut is equivalent to £75 billion of QE*. The model was estimated using historical data since the MPC’s inception in 1997.

The model’s inputs are grouped into indicators of economic growth, inflation and financial market conditions. Growth indicators include the quarterly change in GDP, manufacturing and services purchasing managers’ indices (PMIs) and consumer confidence. The inflation indicators are the deviation of current inflation from target, average earnings growth, consumer inflation expectations and manufacturing firms’ price-raising plans. Financial market indicators include short-term government bond yields, share prices, the effective exchange rate and the interbank / Treasury bill rate spread. The model also takes account of any vote dissents at the prior meeting and inter-meeting signals from individual members (such as Weale’s indication that he will vote to ease in August).

If the MPC changed policy only by moving Bank rate by 25 bp or adding / subtracting to QE in £75 billion chunks, the MPC-ometer would signal action if the forecast average vote moved above +12.5 bp (tightening) or below -12.5 bp (easing). The lower panel of the chart compares the historical fitted values of the model with the implied “unchanged policy range” of +/-12.5 bp, shown shaded. It can be seen that movements outside this range are usually associated with Bank rate changes or – since 2009 – expansions of QE.

The current model forecast for August is -21 bp, the lowest since July 2012, at the height of the Eurozone crisis and during a period of expanding QE. This forecast assumes that tomorrow’s second-quarter GDP release will show quarterly growth of 0.5%, in line with the consensus estimate. The key drivers of the large change this month are the post-referendum slumps in the PMIs and consumer confidence, and a fall in short-term government bond yields.

The forecast is unlikely to change significantly before the MPC vote on 3 August. Apart from the GDP release, a revision could be prompted by final PMI and consumer confidence readings – to the extent that these differ from flash estimates – or movements in markets.

The model is based on average MPC behaviour since its inception. It is possible that the current Committee membership has a different reaction function but there is no evidence that the model has “broken down” in recent years – the stability of policy since 2012 is in line with the average vote forecast, which has remained within the “unchanged” range.

*This assumption maximises the model’s historical fit with MPC behaviour; it does not imply that these alternatives have the same effects on economic growth and inflation.