The November HICP report is slightly better than expected. A step in the right direction for the distributions. The big picture is largely unchanged, but today’s report shows a distribution less dispersed around target and a common component a bit weaker than in recent months. For this reason, we now think it is reasonable to expect MoM (saar) near-term readings in the 2.0-2.5% range. Finally, the medium-term models deliver a forecast almost identical to the time of the flash release; the models forecast remains a bit above target in the medium-term.
We now estimate that the 2024 carryover effect for 2025 is 1 percent for core HICP (1.1% on SA data, 0.8% on NSA data), and 1.3% for headline HICP (1.3% for both, SA and NSA data). Based on this, we discuss the risks around the ECB/NCBs staff forecast, the evolution of our forecast, and whether the 2025 forecast is “too high” at the end of this note.
Evidence from the distributions
The distribution: centered around target. This month, the distribution is less dispersed than last month (ridge plot here). Extending the horizon (Figure 1), the distribution now shows a bit more mass around target in recent months. Finally, the median (Figure 2) has ticked down in November and remains around target.
Overall, this evidence now suggests solid near-term readings a bit above target (MoM saar 2.0-2.5%).
Figure 1. Kernel of HICP excluding food and energy items changes (%, a.r.)
Figure 2. Median of HICP excluding food and energy items prices increase
Evidence from our CI model
Our CI model estimates that net of idiosyncratic shocks, the common component across items is solid. Figure 3 shows the decomposition of the MoM of core HICP in the “common” component and the “idiosyncratic” component. The model estimates that in November the common component increased by 18bps, a bit below the average of the previous months, while the idiosyncratic shock is a small positive (4bps). As we did in previous months, we consider as “true” core the one netting out the idiosyncratic part. Therefore, a rough estimate put the MoM (saar) of “true” core HICP at 2.2% in November. The signal of the CI model in the last few months is in line with the distributions and suggests that core HICP is running a bit above target at the moment. The good news is that the 3m/3m is now at 2.6% and trending down. (see Figure 4 below).
Figure 3. Contributions to MoM changes of HICP excluding food and energy items
Note: the Figure shows the decomposition of the MoM percent changes of HICP prices excluding food and energy items. The contributions are estimated using our CI model, a 2-stage OLS-LASSO regression model.
Figure 4. Estimated “Common” component: YoY, 3m/3m a.r. and 6m/6m a.r.
Note: the Figure shows the 3m/3m at annual rate (green line), the 6m/6m at annual rate (red line), and the YoY (blue line) of the “common component” estimated using our CI model.
Implications for the medium-term forecast of core HICP
Medium-term model-based forecast is unrevised. The models forecasts are essentially unrevised compared to the preview, as the incoming data did not alter the quarterly nowcast. Using the unemployment rate as measure of “slack”, the forecast is at 2.8% (average YoY) in 2024, 2.5% in 2025, 2.3% in 2026, and 2.1% in 2027. Using the output gap, the forecast is: 2.8% in 2024, 2.5% in 2025, 2.3% in 2026, and 2.2% in 2027. The average of these forecasts is a bit above the latest ECB/NCBs staff forecast.
Figure 5. Model-based medium-term forecast of core HICP (YoY)
Using Urate as a measure of “slack”
Using outputp gap as a measure of “slack”
Note: the confidence intervals (C.I.) are calculated using the estimated parameters distributions.
A comparison with the ECB/NCBs staff forecast
Our forecast is persistent, not evolved much in 2024. Some upside risks for 2025. Table 1 shows a comparison between our latest (model-based) forecast and the ECB/NCBs staff forecast. Given that we are at the end of the year, we started to report the 2024 carryover for 2025. For the record, the 2024 carryover for 2025 is similar (just a bit lower) compared to the 2023 carryover for 2024 for core HICP, while it is significantly higher for headline HICP (see here our November 2023 note for a comparison).
Our forecast did not evolve much in 2024. Back in November 2023 (see here), our models forecasted 2.8% for core HICP in 2024, a projection which turned out to be perfect. The forecast did not evolve much during the year, despite some surprises in the MoM readings. For instance, in February the models projected 2.7% for core HICP in 2024 (see here), and went back to 2.8% in June (see here). Overall, we take this as a victory lap, as the ECB/NCB forecast progressively closed the gap with our forecast over the course of the year.
Is the 2025 forecast too high? We often get the question: is the 2025 forecast “too high”? First, we run models and the models estimate a slightly higher persistency than the ECB/NCBs forecast. But the difference is not large. To put some numbers, the average MoM in 2025 consistent with our forecast for core HICP is 20bps, while the one consistent with the ECN/NCBs staff forecast is 17-18bps. In other words, for core HICP the models forecast is well within the range of possible outcomes. As for headline, starting with a 1.3% of carryover, the 2.1% in 2025 of the ECB/NCBs staff implies pretty low readings (16bps on average), not impossible but probably at the low range of the possible outcomes (with the 2.4% of our models being at the upper range of the possible outcomes).
(For a technical note on the concepts of “acquired inflation” and “carryover effect” see here and here).
Table 1. Comparison of forecasts
Note: the “UnderlyingInflation” forecast refers to the average of the two models shown in Figure 5.