In this note we review and discuss the implications for monetary policy of one important chart in today ECB Lane’s speech.
(Preamble not related to this note: Riccardo Trezzi has published an article on VoxEU titled “International policy coordination during disinflation” with Giancarlo Corsetti. The main message of the article is that, so far, inflation has been addressed with a strict national focus. The authors explain why we have not seen cooperation across borders, and under what conditions it might turn out to be useful. Spoiler: the role for cooperation remains very limited and both authors are skeptical we will ever see it in practice during disinflation episodes. “Keep your backyard in order” remains the base case).
The Lane’s Speech
The ECB has published a speech by Philip R. Lane titled “The euro area hiking cycle: an interim assessment”, delivered today at the National Institute of Economic and Social Research in London. For brevity, we have found the (long) speech not particularly informative; therefore, we will not discuss it. However, the speech contains one key information, reviewed below.
Much Higher for Much Longer
The most reliable ECB models suggest that the tightening delivered so far by the ECB is expected to have only a modest impact on inflation, even in case of a recession. Translated: we need much more. Lane’s speech contains two charts summarizing the Impulse Response Functions (IRFs) and expected impact of policy tightening using a set of models maintained by the ECB staff. Specifically, the charts include the IRFs of the ECB-BASE (a large-scale semi-structural model similar to FRB-US), the New Area-Wide Model (NAWM – a fully micro-founded model with a financial accelerator mechanism), the “MMR model” (which is similar to NAWM), the “Basic Model Elasticities” (BME) which is a euro area aggregation of simulation results provided by national central banks, and a Bayesian VAR (BVAR) model.
(Note: for the full list of models maintained in the Eurosystem/ECB, please see here)
The main result is presented in Chart 3 of Lane’s speech which we report below. What you see in the figure is the estimated/expected cumulative impact of policy tightening (that is, the tightening implemented so far) on inflation and GDP growth using each model. The dots show the median across models.
Figure from ECB Lane’s speech
Three main considerations arise from Chart 3 in Lane’s speech:
1) Unsurprisingly, there is a high uncertainty and the estimates appear strictly model-dependent. For instance, the expected disinflationary impact estimated using ECB-BASE in 2025 is ¼ of what NAWM estimates. Therefore, the choice of the model is crucial.
2) NAWM and MMR models estimate a huge (in fact, implausible) impact on GDP growth in 2022-2023. For instance, according to MMR model, without policy tightening GDP growth would have been 4 percentage point higher in 2022, and according to NAWM we should expect a dramatic recession (5.5% GDP contraction) in 2023. Therefore, we can safely disregard both models and interpret with caution the BVAR results for the same reason.
3) The two most reliable models (ECB-BASE and BME) suggest that, conditional on the tightening so far, inflation should moderate only marginally going forward, even in case of a recession. In details, according to ECB-BASE the YoY of inflation (core/headline is irrelevant given the horizon) is expected to drop less than 1 percentage point by 2025, even if GDP growth is expected to drop in 2023 by 2% (which is itself arguable, given the most recent evidence). BME model conveys the same message, although inflation drops even less because output contracts less this year.
Implications for monetary policy
The models imply a terminal rate much higher than 3.5%, kept at that level for a long period of time. As usual, models simulations should be taken with caution, especially now given what happened in the last three years. However, as usual, the models provide the necessary quantitative base for discussion. For instance, looking back, the FRB-US model prediction that a 400bps tightening would have not necessarily resulted in a recession (see Figure 11 here for the output gap FRB-US IRF to a 100bps FF rate shock) turned out to be correct, at least so far. Looking at the euro area, in our view the ECB-BASE and BME models suggest we are in similar spot. Indeed, we start with 5%+ and increasing core HICP inflation (see our latest note here), a stronger than expected cycle, and low sensitivity of output and inflation to an interest rate shock. Taken at face value, the models estimates imply that in order to disinflate the euro area, the ECB would need to (almost) double the tightening put in place so far. We strongly doubt there is any willingness to go that way in Frankfurt but what the models suggest is that ultimately the data might force the ECB to go above 3.5% and stay there for a long period of time.
(Please note that Lane’s speech is much more optimistic and argues that the tightening is expected to reduce inflation by 1.8 percentage points by 2024. This happens because Lane refers to the median estimate across models -the dots in Figure 3 above. The issue however, is that in order to believe to Lane’s point of view one should also believe the NAWM and MMR estimates for output which, as discussed, are unrealistic in our view).