Understanding how markets predict BoC's policy interest rate decisions

Question

I read in the newspaper things like,

Interest rate swaps, which are based on market expectations about future rate decisions, are pricing in at least one Bank of Canada rate cut later this year, and additional cuts in 2024.

Sometimes there will be a probability associated with the prediction, ie. markets are saying that there is 80% chance of decrease to 4.75% the next time the central bank announces its policy interest rate.

My understanding is that there are a few ways of making these predictions, ie using bankers' acceptances or overnight index swaps. Is there a generally accepted belief that one method has greater prediction accuracy than other methods?

Specifically, I'm interested in understanding what and how the markets are predicting for the next Bank of Canada decision for its policy interest rate.

More Questions

• Where exactly do these predictions come from?
• Can they be recreated by publicly available information?
• If so, can someone point me in the right direction as to how to do this?

Answer 1

I trade interest rate derivatives. I can definitively tell you the best way of analysing what is priced in is to identify the liquid and tradeable instruments that most closely aligns with the central bank rate and observe what the price is for that instrument.

In many currencies, GBP, EUR, SEK, NOK, probably USD, there are RFR IRS derivatives that have effective and termination dates that fall exactly between the central bank policy effective dates.

Take Sweden as an example. On Thursday this week (9th Feb 23) they will announce their policy which will apply to the effective dates of Wed 15th Feb 23 to Wed 3rd May 23. The price of the so called "Feb meeting SWESTR" is 2.92%.

The current SWESTR prints consistently at 2.385, when the central bank policy rate is currently 2.50. That means a stable differential of 11.5bps. If you infer the central bank policy rate from 2.92% you get 3.035%, i.e. between the two outcomes of a 50bps hike and (if you assume quarterly point jumps) 75bps then the probabilities are 86% and 14%.