Davide Brignone (ECB), Alistar Dieppe (ECB), and Martino Ricci (ECB) have published on VoxEu an interesting exercise to quantify the risks of persistently higher US inflation.
What the paper does. The analysis uses the ECB-Global 2.0 model to estimate the impact on US inflation of the fiscal stimulus. (for the record: the ECB-Global 2.0 model is a semi-structural, global multi-country model with rich channels of international shock propagation through trade, oil prices and global financial markets). Specifically, the paper constructs three scenarios to quantify the upside risks to inflation considering: (i) a steeper Phillips curve, (ii) stronger fiscal multipliers, and (iii) rising inflation expectations. It should be specified that all simulations are run conditional on the current path of the Federal Funds (that is, the Federal Reserve will not raise interest rates before 2023). All simulations are run in CPI space.
Results. In the baseline, CPI inflation increases by 0.3 percentage points in 2023 as a result of the fiscal stimulus (while it is not particularly well specified by the authors, the Figure below should be interpreted in deviation from the pre-Covid (or at least pre summer 2021) inflation levels, not from the currently elevated readings.). Assuming a steeper Phillips curve (3 times larger slope), the impact on CPI inflation adds an additional seven tenths bringing the total impact to a full percentage point. As for a larger fiscal multiplier, the marginal effect is estimated in between 1 and 2 tenths. Finally, considering the behavior of inflation expectations, their marginal impact is estimated between 2 tenths and 6 tenths according to the scenarios (respectively a moderate increase in long-term inflation expectations and a full de-anchoring). Putting everything together, the authors argue that the upside risks to inflation remain moderate, unless all the risks materialize (in which case, their estimates point to a possible increase of 2.3 percent).
Comment. An interesting exercise because it quantifies the marginal effects of each channel. Overall, I think these estimates can be taken seriously. One limitation of the analysis is that it does not consider PCE inflation. For this reason (and considering that monetary policy does not respond), the authors’ estimates should be seen as upper bounds.
Implications for the Fed Board staff. In our view, this is a mixed bag for the Fed staff. On the one hand, the exercise shows that reaching elevated (4-ish%) readings requires a combination of factors (all of them at the same time) that at the moment seems implausible. However, on the other hand the analysis shows that there is a large number of combinations of these factors that could push realized inflation above the Fed target to a level that several FOMC participants would perceive as uncomfortably high (again, in the baseline CPI inflation rises by 0.3 percentage point in 2023 so any additional effect would be on top of this). The good news for the staff is that the authors’ estimates are conditional to the current SEP median FF rate so, as already mentioned, the Figure below shows an upper bound of the upside risks.