The Bank of England publishes how each Monetary Policy Committee member voted. Tracking the number of dissenting votes across meetings gives a useful forward signal.
Patterns to watch:
A growing minority voting for a change often precedes an actual change by one or two meetings.
Who dissents matters: a shift by the Governor or Deputy Governors carries more weight than external members.
A split in both directions (some for hikes, some for cuts) signals uncertainty.
Market reaction:
A vote split more hawkish or dovish than expected can move GBP more than the decision itself.
The minutes released at the same time explain the reasoning.
Keeping a simple table of votes by member over the last year makes the trend visible.
Do you track MPC votes, and have you seen them anticipate rate changes?
I keep a one page sheet with each MPC member and their last six votes. The useful signal for me is not a single dissent, but the second meeting where the same hawkish minority grows by one name.
Last year a Deputy Governor flipped two meetings before the actual cut, and GBP sold off more on the minutes than on the rate print. External members moving alone rarely changed my bias for more than a session.
I disagree that a split in both directions is always noise. When the Governor sits with the majority and both sides still dissent, I treat that as a pause signal and cut size into the next London open.
The table also stops me reinventing the story after each release. I just add the row and ask whether the minority grew, shrank, or changed who is leading it.
Do you weight the Governor more heavily than a simple headcount? I still wait until three names line up the same way before I lean hard into the next session.
One thing my vote table handles badly is a change of membership. When an external member's term ends, the replacement has no voting history, so for the first two or three meetings that row tells me nothing and I lean on their speeches instead.
Those speeches are an input I log now for everyone. A member who calls policy clearly restrictive in a speech a few weeks before the meeting has, in my notes, voted for a cut at that meeting more often than not.
On the day, the surprise is measured against the expected split rather than against the decision alone. If surveys expect 7 to 2 and the vote comes out 5 to 4 with the same decision, I have seen GBPUSD move more in the first minute than on some actual rate changes.
So I write the expected split on the sheet before 12:00 London, next to my own guess. If I cannot find a consensus split, I assume a repeat of the previous vote and only trade a clear break from it.
How do you handle a meeting where the vote matches expectations but the minutes lean the other way? I usually give the minutes the final word after the first ten minutes.