Both fixing swaps and economist estimates for CPI have shown a tendency to overshoot in recent months.
Yields, though, do not look particularly exposed to a softer-than-expected CPI print today due to their recent fall, and we are unlikely to see a big enough surprise to significantly alter Fed pricing.
I thought it would be interesting to see how accurate CPI fixing swaps are. These fix to year-on-year headline CPI each month, and trade each day. They are traded by inflation dealers, who have a strong monetary incentive to be as accurate as possible.
I took the value of the swap on the day of CPI release and compared it to the actual number. As we can see in the chart below, they are generally pretty accurate, with an average forecast error of 0.06% points. Swaps are more accurate than the median economist estimate, the average of which is 0.09% points.
Interesting to note is that the error of swaps and economists have been much closer in more recent times, suggesting the latter may be paying more attention to fixing swaps when making their estimates.
The chart also shows that both swaps and economists have displayed a positive bias, i.e. they have underestimated CPI, over the last two years (the longest I can currently get fixing swap data), but that bias has switched in the last 6-12 months to overestimating the inflation number.
Thus if this pattern continues, there is some downside risk to today’s number versus expectations. Economists’ median estimate for June’s headline year-on-year CPI is 3.1%, while the latest price of June fixing swap data (as inferred by Bloomberg from inflation swap trades reported to the DTCC) is slightly lower at 3.04%.
A 3% outturn is unlikely to rock the boat as far as risk assets are concerned. The largest downside error from swaps in the last two years is 0.2% points. This too is not enough to alter Fed pricing at this stage, and thus would be unlikely to have a huge impact on bonds and SOFR futures given they have worked off some of their oversoldness in recent days.
