Why US CPI revisions could be a big deal (... or mean nothing at all)

This Friday the Bureau of Labor Statistics will publish the annual update of seasonal adjustment (SA) factors for consumer price inflation, ahead of the January CPI report on 13 February. This will lead to revised MoM% CPI prints for the past five years using those adjustments. Importantly, the non-seasonally data doesn’t change, so the full year YoY% numbers will remain unchanged.

Last year’s revisions were more significant than usual in that they depressed the seasonally adjusted MoM% change for the first half of the year, but it led to upward MoM% revisions for the second half – it is important to remember that seasonal adjustments are a zero sum game over the course of a full year. This indicated that the encouraging moderation in monthly inflation rates through 2022 was no longer so impressive and suggested that the Fed had more work to do to contain price pressures. Rate hike expectations increased, yields rose (contributing to the 2Y Treasury yield rising 60bp over subsequent weeks) and the dollar firmed.

The big changes in 2022 seasonally adjusted month-on-month data were that November became 0.3% MoM from 0.2% and December went to 0.4% MoM from 0.3%. Not huge changes, but these get massively amplified when you look at 3M annualised rates that the Fed have increasingly mentioned as an indicator of inflation momentum. The revisions meant that 3M annualised rate, initially calculated at 3.1% in December 2022, became 4.3% while headline CPI got bumped up from 1.8% to 3.3%! Given the Federal Reserve and markets were caught so off-guard, a couple of Fed officials have said they will carefully be watching them again this year, most notably Fed Governor Christopher Waller. The market has understandably taken note.

As these are purely statistical changes and not reflective of any changes in underlying data, we don’t have a huge amount to go on in determining what will happen on Friday. It is possible that for the second year running it depresses MoM% in first half 2023 and raises it in second half 2023, which would diminish the chances not only of a March cut, but also a May interest rate cut. Conversely we could see seasonally adjusted MoM numbers revised up in first half 2023 and the second half revised down – reversing the story from last year – while leaving the full year 2023 number unchanged. If so that would give the Fed more confidence about inflation pressures dissipating. It would then also bring the story more into line with what the core PCE deflator report is saying – currently running at 1.5% annualised. Of course there is the option that there are no changes whatsoever. We can’t say for certain.

That said the chart below shows the MoM change in the seasonally adjusted core CPI data less the MoM change in the non-seasonally adjusted (NSA) core CPI data. It effectively gives us the seasonal adjustment factors for each month that sum to zero over the whole year. You can see over the past five years the non-seasonally adjusted data (or raw data) post bigger MoM rises in January-April than the seasonally adjusted data while the reverse is the case later in the year, especially November and December.

As you can see the November and December 2023 adjustments above are broadly similar to what we saw for the previous five years. What the revisions did last year was bring 2022 November and December difference between SA & NSA data back into line with history by revising the SA data up by 0.1pp in both months. Given everything for 2023 already looks broadly in line with historical patterns, it perhaps suggests Friday could be a bit of a non-event. That is our base case, but we have to remain open to the potential for a surprise that could meaningfully alter the markets view on the timing of Fed policy changes.

The market is in a growing mood to continue to test higher for longer tenor yield; something we’ve been calling for. If it were to get an excuse to push this a bit more from a revision that was perceived to result in higher annualised inflation readings then some further upward pressure on yields could obtain. It should not be deterministic though, more likely a short-term impulse. The bigger impulse will come from the direction of travel for inflation trends generally. Assuming these continue to point down, then yields can’t rise by too much.

The dramatic rise in the 2yr yield a year ago was not all do the revisions to CPI. We were going through an ongoing upward revision to the terminal peak for the Fed funds rate, and even though CPI inflation was well off the highs, and still falling, it was also still running in the area of 6%, and above the funds rate. There was far less market conviction at the time that the Fed could get inflation back to 2% in a timely fashion.

The 10yr Treasury yield moved from 3.5% to above 4%, which was a large move. But it was arguably too low to begin with. Fast forward to today and the 10yr Treasury yield is at a little over 4%, but practically all inflation metrics are now comfortably below that, leaving less room for re-weights to have a material effect. That said, an inflation friendly re-weighting would have a greater effect than unfriendly ones in our view, as the market is looking for excuses to wind the Fed up for rate cuts. There would still be an effect on unfriendly weightings, but likely less impactful.