Terrific
There are three main takeaways from today’s CPI report: (i) the median of the distribution remains in line with pre-Covid values, (ii) the distribution of price changes continues to send convincing signals of disinflation (although the journey back to the pre-Covid shape is still not obvious or fast), and (iii) the strength of the data (the “common” component of our CI-C model) continues to lose steam.
Overall, the November report confirms the recent signals: there are signs of disinflation. The road to 2% remains uncertain and most likely, it will be bumpy. But the near-term outlook looks very, very promising now.
Evidence from the distributions
This month there are no clear signals from the percentiles of the price changes distribution: some percentiles moved up, some moved down, although they all remained close to the very bening levels of last month. The median (Figure 4) ticked up to 1.2% a.r. (from 0.5% a.r.) but remains in line with the pre-Covid values.
Percentiles details:
- The 5th pct is -30.8% (from -32.3%)
- The 10th pct is -27.2% (from -25.1%)
- The 25th pct is -11.3% (from -11.6%)
- The 50th pct is 1.2% (from 0.5%)
- The 75th pct is 8.5% (from 7.2%)
- The 90th pct is 20.9% (from 18.9%)
- The 95th pct is 27.4% (from 30.5%)
The most important evidence continues to come from the sequential distributions of the last 9 months (Figure 1) which show important signs of stabilization and disinflation. The Kernels of the last 3 months (black line in Figure 1) shows that the distribution is (finally) traveling to the left and it has a thick left shoulder/tail. This is an excellent news.
At the same time, comparing the distribution of the last 3 months to the 2016-2019 period reveals that the road to normalization can be bumpy going forward. Indeed, the most recent distribution (black line) has negative kurtosis (“platykurti”). Put it differently, while pre-Covid the distribution was centered around the Fed target and it had thinner tails (“leptokurtic”), the current distribution has thicker tails (both). Therefore, because the current moderation seems to be driven by the deflation of the left tail, it can be occasionally exposed to the right tail (that is, occasionally we can still have a strong MoM down the road until the right tail will be much thinner).
Evidence from our CI-C model
Our CI-C model estimates that net of Covid and idiosyncratic shocks, the strength of the data in November is losing steam. As such, the CI-C model is confirming the evidence of the distributions.
Figure 5 shows the decomposition of the MoM of CPI excluding food and energy items in the “common” component (the blue bars), the “idiosyncratic” component (the yellow bars), and the “Covid” effect (the green bars). The model estimates that in November the common component increased by 21bps. The Covid effect is estimated nul (0bps), and the idiosyncratic shock is negligible (-2bps).
In other words, the CI-C model suggests that today’s 0.2% in core CPI is genuine because it is not driven by an idiosyncratic shock nor by a Covid-related effect.
Today’s reading brings the YoY of the common component to 2.9 percent, unchanged compared to the previous months. At the same time, today’s reading brings the 3m/3m a.r. of the common component at 2.6 percent and the 6m/6m a.r. at 2.9 percent (Figure 6). This evidence suggests that the YoY of the common component has peaked and it is now ready to start its descent.
Implications for the medium-term forecast of core PCE price inflation
Today’s CPI DOES have material implications for the medium-term profile of core PCE price inflation. Our nowcast of core PCE price inflation in Q4 is now 3.9% (a.r.). Conditional on our Q4 nowcast, our “main” Phillips curve model delivers the following Q4/Q4 forecast: 3.4% in 2023, 3.1% in 2024, and 2.8% in 2025 (see Figure below). The real good news for the Fed is that now the confidence bands of the model include the 2% target at the end of the forecast horizon. We still have a long way, but at least we had our first big positive signal from the medium-term model.
Note: the confidence intervals (C.I.) are estimated using quasi-out-of-sample methods (estimate the model over a sub-sample, forecast, and calculate the root mean squared forecast errors).
First quarter of forecast: 2023:Q1.
Implications for the Fed Board staff
Today’s reading does have meaningful implications for the Fed Board staff.
In 2022 the Fed staff has been -on net- upwardly surprised by the incoming data. However, the Fed staff near-term forecast (as inferred from the FOMC minutes) has been pretty accurate in the last 3 months. In our experience, the staff is unlikely to change materially its near-term forecast for two reasons. First, we are at the end of the year so the near-term right now is much less important (given that the 2022 figure is essentially already determined). Second, rents (notably the most persistent and cyclical item) remained very elevated in November (CPI rent of primary residency came in at 9.6% a.r.).
Putting everything together, in our view the Fed staff will continue to expect core CPI to grow at an average of 30bps in the near future. In our view, the Fed staff near-term forecast balances the risks around it.
Having said so, the estimated model (the “main” Phillips curve model) is finally sending some signs of hope. Today’s downward revision is the first meaningful revision in a year. In our experience, the Fed staff will be careful, given what happened so far. But if anything, the estimated model today moved in the direction of the Fed staff judgmental forecast. As a matter of fact, this is the first time it happens (at least in this magnitude) since the beginning of this inflation episode. As such, this is a signal that the reader should not ignore. In our view, this is still not enough for the FOMC to pivot. But this time Powell, the Fed staff and the FOMC are taking note.
Note: given the timing of the FOMC, we do not expect the Fed staff to submit updated medium-term forecasts (not enough time). Rather, in our experience the Fed staff will submit an updated near-term forecast and a general comment, typically delivered by a senior officer.
Figures – monthly distributions
Figure 2. Percentiles and Standard Deviation of the distribution of MoM changes (CPI prices excluding food and energy items, % a.r.)
Figures – CI-C model
Figure 5. Contributions to MoM changes of CPI excluding food and energy items (CI-C model)
Note: the Figure shows the decomposition of the MoM percent changes of CPI prices excluding food and energy items. The contributions are estimated using our CI-C model, a 2-stage OLS-LASSO regression model. The “Covid” effect is identified with price variations outside the 10th-90th percentiles of each item pre-Covid price change distribution.