Riccardo Trezzi (University of Geneva, UnderlyingInflation) has published a paper titled “The euro area great inflation surge” together with Guido Ascari (Oxford, Pavia, Dutch National Bank (DNB)), Paolo Bonomolo (DNB), and Marco Hoeberichts (DNB). The paper extends an analysis we circulated back in December (see here). In the minds of the authors (original interpretation), the paper is meant to re-write the dominant narrative of EA inflation (“it is only/mainly driven by an energy shock”). The conclusion is that demand must be much stronger than commonly believed. Therefore, without some help from fiscal policy, the ECB has a very, very tough job ahead.
The intuition
It cannot only (or mainly) be driven by energy prices. The authors suggest a narrative of the Euro Area inflation surge, supported by theoretical (AS-AD) analysis and empirical (Bayesian VAR) evidence. The authors show why the current debate, often solely focusing on energy prices, has failed to appreciate the complexity of what happened: demand and supply both contributed to output and inflation dynamics and their absolute (and relative) contributions have changed over time. The main result of the paper is that while price pressures were already building before the natural gas shock, an inflationary impulse in one sector (energy) has quickly become broad based because demand conditions remained accommodative facing the negative supply shock. Put it simply: it is about demand, not (only) supply. Policy implications follow and imply a big effort of both monetary and fiscal.
What happened through the lengths of a Macro101 AS-AD model
Supply and demand both contributed. The authors define three chronological phases. Phase I: the covid shock (Q1 and Q2:2020); Phase II: the re-opening of the economy (Q3:2020 to Q3:2021) Phase III: the post re-opening (Q4:2021 to present). With the lens of a simple aggregate demand and aggregate supply (AS-AD) model, the authors first set the narrative of the relative movements of inflation and real activity of the last 3 years. Figure 1 shows the evolution of core HICP inflation and real output in the Euro area. Figure 2 shows the authors’ explanation of what happened in each phase using the AS-AD diagram.
Figure 1. Core HICP inflation rate and real GDP in the Euro area
Figure 2. AD-AS diagram and the three phases.
The Covid shock (Phase I) and the following restrictions are usually interpreted as a combination of a negative supply shock and a negative demand shock. Indeed, Euro Area GDP shows an unprecedented fall and a deceleration of inflation, suggesting that in phase I the large negative demand shock more than offset the negative supply shock. In the AS-AD diagram, aggregate supply (S) shifts up and to the left (from S to S’), and aggregate demand (D) shifts down to the left (from D to D’).
The re-opening of the economy (Phase II) was characterized by an increase in demand and supply constraints. As a result, output expanded and inflation increased. Aggregate demand, both private and public (stimuli), expanded sharply in phase II. In some dimensions (i.e. retail sales), demand went above pre-Covid levels. Aggregate supply experienced difficulties in copying with the increase in demand, because of global supply chains bottlenecks, intermediate goods and labor shortages. In terms of the AS-AD diagram, the economy moved from E’ to E’’, a point associated with output drifting to pre-Covid levels, but with higher inflation.
The post-reopening phase experienced an additional negative supply shock and a positive aggregate demand shock, resulting in further price pressures. In phase III, global supply chains recovered and the economies reopened further but starting in late 2021, aggregate supply in the Euro Area was hit again by a negative shock, although of a different kind this time: a sharp increase in energy prices. Both households and firms raised their consumption/investment in mid-2022 as a result of pent-up demand and extra savings. The labor market remained tight and government consumption remained elevated. In terms of the AS-AD model, the economy moved from E’’ to E’’’, a point with output similar to pre-Covid but with much higher inflation.
To sum up, as shown in Table 1 below, in each phase both demand and supply have moved. According to the authors, this is why the complexity of the EA inflation phenomenon is generally underappreciated. It also means that the widespread claim that “EA inflation is mainly/only supply-side / energy-driven” is wrong. You cannot have inflation without a recession if the only/mainly shock is a negative supply shock: demand must play an even bigger role. The issue is about demand, not supply.
What happened through the lengths of a Bayesian VAR
A Bayesian VAR confirms that in the data demand factors are (at least) as important as supply-side shocks, as in the AS-AD diagram. Without entering in the details for brevity, the authors identify the contributions of demand and supply side factors (distinguishing between gas/energy and other supply-side shocks) using the forecast errors of a BVAR identified with signs restrictions. Figure 3 shows the authors’ results. The blue bars show the contributions of supply-side factors, the light blue bars those of demand-side factors, and the yellow bars the contribution of energy (the black line shows the BVAR forecast errors). The bottom line is that the BVAR confirms the absolute and relative contributions of demand and supply in the 3 phases. In phase I the negative demand shock more than offset the negative supply shock and inflation surprised to the downside. In phase II the positive demand shock offset the (partial) positive supply shock and inflation surprised to the upside. Finally, in phase III inflation surprised to the upside as a result of positive contributions of the three of them: demand, supply, and energy prices.
Figure 3. VAR-based forecast errors (output and inflation)
Note: the figure shows the forecast errors decomposition of a BVAR. The left column refers to output, while the right column refers to inflation. The dark blue bars (“AS”) show the contributions of aggregate supply factors, the light blue bars (“AD”) those of aggregate demand, and the yellow bars (“NRG”) the contribution of energy shocks.
Conclusion
Monetary policy is in trouble, might need fiscal as well. If the authors are right, the ECB is in trouble because the dominant narrative (“inflation in the EA is mainly/only energy driven”) is wrong: accommodative demand has played a crucial role in this story. Reducing an inflationary process that is now centered around a mean/median of about 6% (and still growing) is no joke. If we take the ECB models prediction at face value, 4% terminal rate would not be enough without help from fiscal policy. The good news is that we are now sure someone in the Eurosystem knows. The hope is that the paper will be read in Frankfurt.