Large Idiosyncratic Shock, Distribution around Target
The November HICP report brought some good news. The good news is that the distribution of price changes has shifted to the left in the last 3 months, and that the median of the distribution has dropped to (or just below) target. The CI-C model suggests that the weak November reading is mostly driven by a large idiosyncratic shock in one sector that is unlikely to materialize again going forward. Having said that, given the distribution, it is now reasonable to expect moderate core HICP readings going forward (i.e. MoM saar around 2.5%). Finally, the medium-term models are little changed and the risks around the new ECB forecast are well balanced.
Going forward, we expect the big picture to remain broadly unchanged at least until the end of January when a new quarter (Q1) will be put in-sample in the models. Until then, in our view, we will deal with a central bank that cannot admit to be / turn dovish because the medium-term forecast does not converge to target even if the near-term data are weakening.
Evidence from the distributions
The distribution: very good progress. This month, all percentiles moved down (ridge plot here), confirming the signals of the last few months. Not only, but in the last 3 months (the black line in Figure 1) the distribution has shifted to the left, an important sign of disinflation. Finally, the median (Figure 2) has dropped notably compared to a year ago and it is now a touch below target.
Overall, this evidence is now solid. The distribution suggests near-term readings around target (MoM saar around 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-C model
Our CI-C model estimates that net of Covid and idiosyncratic shocks, the common component across items in November was in line with previous months. Figure 3 shows the decomposition of the MoM of core HICP in the “common” component, the “idiosyncratic” component, and the “Covid” effect. The model estimates that in November the common component increased by 17bps, in line with the average of the previous months. The Covid effect is estimated at 3bps, and the idiosyncratic shock is large negative (-9bps). 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 around 2.5% in November, a bit below the previous months. The signal of the CI-C model is now in line with the distributions.
(Note: According to the CI-C model, the November reading was driven by a large negative idiosyncratic shock in the services related to transport sector (see unchained NSA level here). As such, the shock is unlikely to materialize again going forward).
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-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.
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-C model.
Implications for the medium-term forecast of core HICP
Medium-term model-based forecast is unchanged. Today’s data imply only cosmetic revisions to the medium-term model-based forecast, as the final reading for November was in line with the flash estimate. Using the unemployment rate as measure of “slack”, the forecast is at 2.9% (average YoY) in 2024, 2.6% in 2025, and 2.5% in 2026. Using the output gap (right panel in Figure 5), the model delivers a more dovish forecast: 2.7% in 2024, 2.1% in 2025, and 2.1% in 2026. See below for a comparison with the 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.
First quarter of forecast: 2024:Q1.
A comparison with the ECB/NCBs staff forecast
Risks around the ECB/NCBs forecasts are balanced. Table 1 shows a comparison between our updated forecast and the new ECB/NCBs staff forecast. The 2023 carryover for 2024 is 1.2%-1.3%. This implies that a forecast a bit below 3% for 2024 is reasonable. Going beyond 2024, the models are a touch above the ECB/NCBs staff forecast, as the inflation process is estimated a bit more persistent. Nevertheless, the gap is now very small. Risks around this forecast are very well balanced.
(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.
Implications for the ECB
No big implications for the ECB. We just had the December Governing Council meeting. We point the reader to our GC notes (here and here). We only add that the impression is that the forecast was set wisely this time and that the Governing Council will have no reason to sound more dovish (or less dovish) until March. The models are unlikely to revise again their forecast until Q1 (preliminary January HICP reading). At the end of January, we will revisit the implications for the ECB. In the meanwhile, we continue to expect moderate core HICP readings, as explained above.