TheMotorcycleBoy wrote:dealtn wrote:A noticeable rise in reported inflation day, and the trigger for the Bank of England to explain its thinking.
https://www.bbc.co.uk/news/business-58563417Whilst the rate isn't that far above the top of the Bank's target, nor significantly high by historical standards, it does mark the steepest monthly increase since comparable records began in 1997.
So that's (3.2%) about 0.6% less than the value predicted by Gilts vs ILGs (3.8ish%).
Help me out and show how me to derive inflation forecasts from the inflation swap curves.
Matt
Firstly the Index Linked Gilts all reference RPI, not CPI. RPI is 4.8% according to todays announcement.
Secondly the breakeven calculation from the Gilts market is complicated. I didn't read your BofE numbers from the other day but I think they were 3.5%ish across the curve. What's important to note though is this is a "spot" curve not a "forward" curve so it isn't intuitively obvious what the predicted rate of inflation is.
In simple terms if 1 year "spot" inflation is priced at 3.5% and 2 year inflation is also priced at 3.5%, then the 1 year inflation rate in a years time is also priced at 3.5%. The 2 year rate is the average, and since we know the 1 year rate is also 3.5%, then the second year must also be 3.5%.
Were spot prices such that 1 year was 2% and 2 year was 3% then we can infer that (approximately) the 1 year rate starting in 1 years time is 4% (since 1 year at 2% followed by 1 year at 4% gives a 2 year average of 3%). This is called "bootstrapping" where you can construct the forward implied rate (be that inflation, or interest, or FX etc.) from all known spot prices along a curve to derive successive unknown future points in between.
In the example above we knew the 1 year and 2 year rates to get the unknown 1 year rate starting in 1 years time. With a 3 year spot rate (and a 1 year and 2 year and the 1 year in 1 years time rate) we can calculate the 1 year rate starting in 2 years time. With a 4 year (knowing the 1 year, 2 year, 3 year, 1 year 1 year forward, 1 year 2 years forward) you can know calculate the 1 year 3 years forward rate etc.
In practice you will have a pricing model more complicated than this and can derive all sorts of more complicated future rates of multiple periods, and create zero coupon real and nominal interest rates for any future date to use to discount any predicted cashflows back to their net present values (NPV) etc.
I hope I didn't lose you.